{ "name": "AI release timeline", "publisher": "Holistic Hybrid Inc.", "url": "https://guidetoartificialintelligence.com/en/ai-timeline/", "license": "CC BY 4.0", "version": "1.0", "updated": "2026-10-26", "events": [ { "date": "-0750-01-01", "date_label": "c. 8th century BCE", "lab": "history", "type": "myth", "title": "Homer describes Hephaestus' golden attendants", "summary": "The Iliad describes the smith god Hephaestus working with golden maidservants who can move, speak and think, one of the oldest images of artificial helpers.", "primary_source": "https://www.perseus.tufts.edu/hopper/text?doc=Perseus:text:1999.01.0134:book=18", "secondary_source": "https://en.wikipedia.org/wiki/Hephaestus" }, { "date": "-0250-01-01", "date_label": "c. 3rd century BCE", "lab": "history", "type": "myth", "title": "Talos, the bronze guardian of Crete, in the Argonautica", "summary": "Apollonius of Rhodes' Argonautica tells of Talos, a bronze giant made to guard Crete, who is defeated when the fluid in his single vein drains away.", "primary_source": "https://www.perseus.tufts.edu/hopper/text?doc=Perseus:text:1999.01.0228:book=4", "secondary_source": "https://en.wikipedia.org/wiki/Talos" }, { "date": "0008-01-01", "date_label": "c. 8 CE", "lab": "history", "type": "literature", "title": "Ovid's Pygmalion: a statue brought to life", "summary": "In the Metamorphoses, Ovid tells of the sculptor Pygmalion, whose ivory statue is brought to life, an early story of a made figure becoming a person.", "primary_source": "", "secondary_source": "https://en.wikipedia.org/wiki/Pygmalion_(mythology)" }, { "date": "0060-01-01", "date_label": "1st century CE", "lab": "history", "type": "invention", "title": "Hero of Alexandria writes on automata", "summary": "Hero of Alexandria described programmable automata and self-moving theatre devices driven by weights, ropes and pegged drums.", "primary_source": "", "secondary_source": "https://en.wikipedia.org/wiki/Hero_of_Alexandria" }, { "date": "1206-01-01", "date_label": "1206", "lab": "history", "type": "invention", "title": "Al-Jazari's Book of Knowledge of Ingenious Mechanical Devices", "summary": "The engineer Ismail al-Jazari completed an illustrated treatise describing water clocks and automata, including mechanical musicians and a serving figure.", "primary_source": "", "secondary_source": "https://en.wikipedia.org/wiki/Ismail_al-Jazari" }, { "date": "1305-01-01", "date_label": "c. 1305", "lab": "history", "type": "concept", "title": "Ramon Llull's Ars Magna", "summary": "The Majorcan philosopher Ramon Llull presented a method of combining basic concepts with rotating paper discs to generate statements mechanically, later cited as an ancestor of logical computation.", "primary_source": "https://plato.stanford.edu/entries/llull/", "secondary_source": "https://en.wikipedia.org/wiki/Ramon_Llull" }, { "date": "1495-01-01", "date_label": "c. 1495", "lab": "history", "type": "invention", "title": "Leonardo da Vinci's mechanical knight", "summary": "Leonardo da Vinci sketched an armoured knight moved by pulleys and cables that could sit, stand and move its arms.", "primary_source": "", "secondary_source": "https://en.wikipedia.org/wiki/Leonardo%27s_robot" }, { "date": "1580-01-01", "date_label": "16th century (legend)", "lab": "history", "type": "myth", "title": "The golem of Prague legend", "summary": "Jewish folklore tells of a golem, a clay figure animated through sacred words; the best-known version credits Rabbi Judah Loew of Prague with making one to protect the community.", "primary_source": "", "secondary_source": "https://en.wikipedia.org/wiki/Golem" }, { "date": "1637-01-01", "date_label": "1637", "lab": "history", "type": "concept", "title": "Descartes asks whether machines could think", "summary": "In the Discourse on the Method, René Descartes argued that a machine could imitate a body but could never use language flexibly or reason across all situations as a person does.", "primary_source": "https://www.gutenberg.org/ebooks/59", "secondary_source": "https://plato.stanford.edu/entries/descartes/" }, { "date": "1651-01-01", "date_label": "1651", "lab": "history", "type": "concept", "title": "Hobbes: reasoning is reckoning", "summary": "In Leviathan, Thomas Hobbes described reasoning as a form of calculation, adding and subtracting ideas, a view later linked to computational theories of mind.", "primary_source": "https://www.gutenberg.org/ebooks/3207", "secondary_source": "https://en.wikipedia.org/wiki/Leviathan_(Hobbes_book)" }, { "date": "1666-01-01", "date_label": "1666", "lab": "history", "type": "concept", "title": "Leibniz imagines a calculus of reasoning", "summary": "Gottfried Wilhelm Leibniz's Dissertation on the Art of Combinations began his project of a universal symbolic language and a calculus ratiocinator in which disputes could be settled by calculation.", "primary_source": "https://plato.stanford.edu/entries/leibniz-logic-influence/", "secondary_source": "https://en.wikipedia.org/wiki/Calculus_ratiocinator" }, { "date": "1726-10-28", "date_label": "28 October 1726", "lab": "history", "type": "literature", "title": "Swift's 'Engine' in Gulliver's Travels", "summary": "Jonathan Swift's satire describes a professor at Lagado whose engine of rotating word blocks produces books by random combination, an early parody of mechanical text generation.", "primary_source": "https://www.gutenberg.org/ebooks/829", "secondary_source": "https://en.wikipedia.org/wiki/The_Engine" }, { "date": "1739-01-01", "date_label": "1739", "lab": "history", "type": "invention", "title": "Vaucanson's Digesting Duck", "summary": "Jacques de Vaucanson exhibited a mechanical duck that appeared to eat and digest grain, one of the most famous automata of the 18th century.", "primary_source": "", "secondary_source": "https://en.wikipedia.org/wiki/Digesting_Duck" }, { "date": "1768-01-01", "date_label": "1768–1774", "lab": "history", "type": "invention", "title": "Jaquet-Droz automata: the Writer, the Draughtsman and the Musician", "summary": "Swiss watchmakers Pierre Jaquet-Droz, his son Henri-Louis and Jean-Frédéric Leschot built three automata, including a boy who writes a programmable text with a quill.", "primary_source": "", "secondary_source": "https://en.wikipedia.org/wiki/Jaquet-Droz_automata" }, { "date": "1770-01-01", "date_label": "1770", "lab": "history", "type": "invention", "title": "The Mechanical Turk chess automaton", "summary": "Wolfgang von Kempelen presented a chess-playing automaton at the Habsburg court; it toured for decades before being revealed as a hidden human operator.", "primary_source": "", "secondary_source": "https://en.wikipedia.org/wiki/Mechanical_Turk" }, { "date": "1816-01-01", "date_label": "1816", "lab": "history", "type": "literature", "title": "E. T. A. Hoffmann's 'The Sandman'", "summary": "Hoffmann's story follows a student who falls in love with Olimpia, a lifelike automaton, a theme later studied as the uncanny.", "primary_source": "", "secondary_source": "https://en.wikipedia.org/wiki/The_Sandman_(short_story)" }, { "date": "1818-01-01", "date_label": "1 January 1818", "lab": "history", "type": "literature", "title": "Mary Shelley's Frankenstein", "summary": "Mary Shelley's novel tells of a scientist who creates a living being and then abandons it, a lasting reference point for debates about creators and their creations.", "primary_source": "https://www.gutenberg.org/ebooks/41445", "secondary_source": "https://en.wikipedia.org/wiki/Frankenstein" }, { "date": "1822-01-01", "date_label": "1822", "lab": "history", "type": "invention", "title": "Babbage proposes the Difference Engine", "summary": "Charles Babbage announced plans for a machine to calculate mathematical tables automatically, the first of his mechanical computing designs.", "primary_source": "https://www.computerhistory.org/babbage/", "secondary_source": "https://en.wikipedia.org/wiki/Difference_engine" }, { "date": "1837-01-01", "date_label": "1837", "lab": "history", "type": "invention", "title": "Babbage describes the Analytical Engine", "summary": "Babbage designed a general-purpose mechanical computer with a store, a mill and punched-card input, though it was never completed.", "primary_source": "", "secondary_source": "https://en.wikipedia.org/wiki/Analytical_engine" }, { "date": "1843-01-01", "date_label": "1843", "lab": "history", "type": "paper", "title": "Ada Lovelace's notes on the Analytical Engine", "summary": "Ada Lovelace translated Menabrea's article on the Analytical Engine and added notes, including a method to compute Bernoulli numbers and the remark that the machine could not originate anything on its own.", "primary_source": "https://www.fourmilab.ch/babbage/sketch.html", "secondary_source": "https://en.wikipedia.org/wiki/Ada_Lovelace" }, { "date": "1854-01-01", "date_label": "1854", "lab": "history", "type": "paper", "title": "Boole's The Laws of Thought", "summary": "George Boole published an algebra of logic in which reasoning is expressed with symbols and operations, the basis of the Boolean logic used in computers.", "primary_source": "https://www.gutenberg.org/ebooks/15114", "secondary_source": "https://plato.stanford.edu/entries/boole/" }, { "date": "1863-06-13", "date_label": "13 June 1863", "lab": "history", "type": "literature", "title": "Samuel Butler's 'Darwin among the Machines'", "summary": "In a letter to a New Zealand newspaper, Samuel Butler suggested that machines might evolve and one day overtake humanity.", "primary_source": "", "secondary_source": "https://en.wikipedia.org/wiki/Darwin_among_the_Machines" }, { "date": "1872-01-01", "date_label": "1872", "lab": "history", "type": "literature", "title": "Butler's Erewhon and 'The Book of the Machines'", "summary": "Butler's novel Erewhon describes a society that has banned machines for fear they would develop consciousness.", "primary_source": "https://www.gutenberg.org/ebooks/1906", "secondary_source": "https://en.wikipedia.org/wiki/Erewhon" }, { "date": "1886-01-01", "date_label": "1886", "lab": "history", "type": "literature", "title": "L'Ève future popularises the word 'android'", "summary": "Auguste Villiers de l'Isle-Adam's novel features an artificial woman built by a fictional Thomas Edison and helped spread the term andréide (android).", "primary_source": "", "secondary_source": "https://en.wikipedia.org/wiki/The_Future_Eve" }, { "date": "1909-11-01", "date_label": "November 1909", "lab": "history", "type": "literature", "title": "E. M. Forster's 'The Machine Stops'", "summary": "Forster's short story imagines people living isolated lives entirely dependent on a global machine that eventually fails.", "primary_source": "", "secondary_source": "https://en.wikipedia.org/wiki/The_Machine_Stops" }, { "date": "1912-01-01", "date_label": "1912", "lab": "history", "type": "invention", "title": "Torres Quevedo's El Ajedrecista", "summary": "Spanish engineer Leonardo Torres Quevedo built an electromechanical machine that could play a king-and-rook chess endgame against a human without help.", "primary_source": "", "secondary_source": "https://en.wikipedia.org/wiki/El_Ajedrecista" }, { "date": "1921-01-25", "date_label": "25 January 1921", "lab": "history", "type": "literature", "title": "Karel Čapek's R.U.R. introduces the word 'robot'", "summary": "Čapek's play about artificial workers who rebel against their makers premiered in Prague and gave the world the word robot.", "primary_source": "https://www.gutenberg.org/ebooks/59112", "secondary_source": "https://en.wikipedia.org/wiki/R.U.R." }, { "date": "1927-01-10", "date_label": "10 January 1927", "lab": "history", "type": "film", "title": "Fritz Lang's Metropolis", "summary": "Lang's silent film, which features a robot made to impersonate a woman named Maria, premiered in Berlin and shaped later screen images of robots.", "primary_source": "", "secondary_source": "https://en.wikipedia.org/wiki/Metropolis_(1927_film)" }, { "date": "1936-01-01", "date_label": "1936", "lab": "history", "type": "paper", "title": "Turing's 'On Computable Numbers'", "summary": "Alan Turing described an abstract machine that manipulates symbols on a tape and showed that a universal machine could carry out any computation, a foundation of computer science.", "primary_source": "https://www.cs.virginia.edu/~robins/Turing_Paper_1936.pdf", "secondary_source": "https://plato.stanford.edu/entries/turing-machine/" }, { "date": "1942-03-01", "date_label": "March 1942", "lab": "history", "type": "literature", "title": "Asimov's Three Laws of Robotics in 'Runaround'", "summary": "Isaac Asimov's short story 'Runaround' set out the Three Laws of Robotics, rules meant to keep robots safe and obedient to humans.", "primary_source": "", "secondary_source": "https://en.wikipedia.org/wiki/Three_Laws_of_Robotics" }, { "date": "1943-01-01", "date_label": "1943", "lab": "history", "type": "paper", "title": "McCulloch and Pitts model the neuron", "summary": "Warren McCulloch and Walter Pitts showed that networks of simple on/off model neurons could compute logical functions, the first mathematical model of a neural network.", "primary_source": "https://doi.org/10.1007/BF02478259", "secondary_source": "https://en.wikipedia.org/wiki/Artificial_neuron" }, { "date": "1945-07-01", "date_label": "July 1945", "lab": "history", "type": "paper", "title": "Vannevar Bush's 'As We May Think' imagines the memex", "summary": "In The Atlantic Monthly, Vannevar Bush described the memex, a desk that would store a person's documents on microfilm and link any two items into trails, an early vision of hypertext and personal information machines.", "primary_source": "https://www.theatlantic.com/magazine/archive/1945/07/as-we-may-think/303881/", "secondary_source": "https://en.wikipedia.org/wiki/As_We_May_Think" }, { "date": "1946-02-01", "date_label": "February 1946", "lab": "history", "type": "invention", "title": "ENIAC is unveiled", "summary": "The University of Pennsylvania publicly presented ENIAC, a programmable electronic general-purpose computer built for the U.S. Army.", "primary_source": "", "secondary_source": "https://en.wikipedia.org/wiki/ENIAC" }, { "date": "1948-01-01", "date_label": "1948", "lab": "history", "type": "paper", "title": "Norbert Wiener's Cybernetics", "summary": "Norbert Wiener's book introduced cybernetics, the study of control and communication through feedback in animals and machines.", "primary_source": "", "secondary_source": "https://en.wikipedia.org/wiki/Cybernetics:_Or_Control_and_Communication_in_the_Animal_and_the_Machine" }, { "date": "1948-01-01", "date_label": "1948", "lab": "history", "type": "paper", "title": "Turing's report 'Intelligent Machinery'", "summary": "In an unpublished report for the National Physical Laboratory, Turing discussed machines that learn and early network-like 'unorganised machines'.", "primary_source": "", "secondary_source": "https://en.wikipedia.org/wiki/Unorganized_machine" }, { "date": "1949-01-01", "date_label": "1949", "lab": "history", "type": "paper", "title": "Hebb's The Organization of Behavior", "summary": "Psychologist Donald Hebb proposed that connections between neurons strengthen when they are active together, a learning rule that influenced neural network research.", "primary_source": "", "secondary_source": "https://en.wikipedia.org/wiki/Hebbian_theory" }, { "date": "1949-01-01", "date_label": "1948–1949", "lab": "history", "type": "invention", "title": "Grey Walter's robot tortoises", "summary": "Neurophysiologist William Grey Walter built small autonomous robots, Elmer and Elsie, that used light sensors to steer and find their charging station.", "primary_source": "", "secondary_source": "https://en.wikipedia.org/wiki/William_Grey_Walter" }, { "date": "1950-03-01", "date_label": "March 1950", "lab": "history", "type": "paper", "title": "Shannon's 'Programming a Computer for Playing Chess'", "summary": "Claude Shannon set out how a computer could play chess by evaluating positions and searching ahead, a template for game-playing programs.", "primary_source": "https://doi.org/10.1080/14786445008521796", "secondary_source": "https://en.wikipedia.org/wiki/Claude_Shannon" }, { "date": "1950-10-01", "date_label": "October 1950", "lab": "history", "type": "paper", "title": "Turing's 'Computing Machinery and Intelligence'", "summary": "Turing replaced the question 'Can machines think?' with the imitation game, now called the Turing test, and answered common objections to machine intelligence.", "primary_source": "https://doi.org/10.1093/mind/LIX.236.433", "secondary_source": "https://plato.stanford.edu/entries/turing-test/" }, { "date": "1950-12-02", "date_label": "2 December 1950", "lab": "history", "type": "literature", "title": "Asimov's I, Robot", "summary": "Asimov collected his robot stories, built around the Three Laws, into the book I, Robot.", "primary_source": "", "secondary_source": "https://en.wikipedia.org/wiki/I,_Robot" }, { "date": "1951-01-01", "date_label": "1951", "lab": "history", "type": "invention", "title": "Minsky and Edmonds build SNARC", "summary": "Marvin Minsky and Dean Edmonds built SNARC, a machine of about 40 artificial neurons using vacuum tubes that simulated a rat learning a maze.", "primary_source": "", "secondary_source": "https://en.wikipedia.org/wiki/Stochastic_Neural_Analog_Reinforcement_Calculator" }, { "date": "1952-01-01", "date_label": "1951–1952", "lab": "history", "type": "program", "title": "Strachey's draughts program", "summary": "Christopher Strachey wrote one of the first game-playing programs, a draughts (checkers) player that ran on the Ferranti Mark 1.", "primary_source": "", "secondary_source": "https://en.wikipedia.org/wiki/Christopher_Strachey" }, { "date": "1954-01-07", "date_label": "7 January 1954", "lab": "history", "type": "event", "title": "Georgetown–IBM machine translation demonstration", "summary": "Georgetown University and IBM publicly demonstrated a computer translating more than sixty Russian sentences into English, the first public machine-translation demo.", "primary_source": "", "secondary_source": "https://en.wikipedia.org/wiki/Georgetown%E2%80%93IBM_experiment" }, { "date": "1955-08-31", "date_label": "31 August 1955", "lab": "history", "type": "paper", "title": "Dartmouth proposal coins 'artificial intelligence'", "summary": "John McCarthy, Marvin Minsky, Nathaniel Rochester and Claude Shannon proposed a summer research project on 'artificial intelligence', the first known use of the term.", "primary_source": "http://jmc.stanford.edu/articles/dartmouth/dartmouth.pdf", "secondary_source": "https://en.wikipedia.org/wiki/Dartmouth_workshop" }, { "date": "1956-01-01", "date_label": "1956", "lab": "history", "type": "program", "title": "Newell, Shaw and Simon's Logic Theorist", "summary": "Allen Newell, Cliff Shaw and Herbert Simon presented the Logic Theorist, a program that proved theorems from Principia Mathematica, often called the first AI program.", "primary_source": "", "secondary_source": "https://en.wikipedia.org/wiki/Logic_Theorist" }, { "date": "1956-06-01", "date_label": "Summer 1956", "lab": "history", "type": "event", "title": "The Dartmouth Summer Research Project", "summary": "About a dozen researchers met at Dartmouth College over roughly eight weeks; the workshop is widely treated as the founding event of AI as a field.", "primary_source": "https://home.dartmouth.edu/about/artificial-intelligence-ai-coined-dartmouth", "secondary_source": "https://en.wikipedia.org/wiki/Dartmouth_workshop" }, { "date": "1957-01-01", "date_label": "1957", "lab": "history", "type": "program", "title": "General Problem Solver", "summary": "Newell, Shaw and Simon began the General Problem Solver, a program meant to solve many kinds of formal problems using means-ends analysis.", "primary_source": "", "secondary_source": "https://en.wikipedia.org/wiki/General_Problem_Solver" }, { "date": "1958-01-01", "date_label": "1958", "lab": "history", "type": "research", "title": "Rosenblatt's Perceptron", "summary": "Frank Rosenblatt published the perceptron, a trainable single-layer neural network, and demonstrated it for the U.S. Navy; the Mark I Perceptron hardware followed.", "primary_source": "https://doi.org/10.1037/h0042519", "secondary_source": "https://en.wikipedia.org/wiki/Perceptron" }, { "date": "1958-01-01", "date_label": "1958", "lab": "history", "type": "invention", "title": "John McCarthy creates Lisp", "summary": "McCarthy designed Lisp, a programming language based on symbolic lists that became the main language of AI research for decades.", "primary_source": "http://www-formal.stanford.edu/jmc/recursive.pdf", "secondary_source": "https://en.wikipedia.org/wiki/Lisp_(programming_language)" }, { "date": "1959-01-01", "date_label": "1959", "lab": "history", "type": "event", "title": "McCarthy and Minsky start the MIT AI Project", "summary": "McCarthy and Minsky founded the Artificial Intelligence Project at MIT, which grew into one of the field's leading laboratories.", "primary_source": "", "secondary_source": "https://en.wikipedia.org/wiki/MIT_Computer_Science_and_Artificial_Intelligence_Laboratory" }, { "date": "1959-07-01", "date_label": "July 1959", "lab": "history", "type": "paper", "title": "Arthur Samuel's checkers program and 'machine learning'", "summary": "Arthur Samuel of IBM published his work on a checkers program that improved by playing itself, helping establish the term machine learning.", "primary_source": "https://doi.org/10.1147/rd.33.0210", "secondary_source": "https://en.wikipedia.org/wiki/Arthur_Samuel_(computer_scientist)" }, { "date": "1961-01-01", "date_label": "1961", "lab": "history", "type": "invention", "title": "Unimate, the first industrial robot, goes to work", "summary": "The Unimate robotic arm, developed by George Devol and Joseph Engelberger, began handling hot die-cast parts at a General Motors plant in New Jersey.", "primary_source": "", "secondary_source": "https://en.wikipedia.org/wiki/Unimate" }, { "date": "1965-01-01", "date_label": "1965", "lab": "history", "type": "program", "title": "DENDRAL project begins", "summary": "Edward Feigenbaum, Joshua Lederberg and colleagues at Stanford started DENDRAL, a program that inferred molecular structures from mass-spectrometry data and became the first expert system.", "primary_source": "", "secondary_source": "https://en.wikipedia.org/wiki/Dendral" }, { "date": "1966-01-01", "date_label": "January 1966", "lab": "history", "type": "program", "title": "Weizenbaum's ELIZA", "summary": "Joseph Weizenbaum published ELIZA, a program that held text conversations by matching patterns; its 'DOCTOR' script imitated a psychotherapist and many users attributed understanding to it.", "primary_source": "https://doi.org/10.1145/365153.365168", "secondary_source": "https://en.wikipedia.org/wiki/ELIZA" }, { "date": "1966-01-01", "date_label": "1966–1972", "lab": "history", "type": "research", "title": "Shakey the robot at SRI", "summary": "SRI's Shakey was the first mobile robot to reason about its own actions, combining vision, planning and movement; the project produced the A* search algorithm.", "primary_source": "https://www.sri.com/hoi/shakey-the-robot/", "secondary_source": "https://en.wikipedia.org/wiki/Shakey_the_robot" }, { "date": "1966-11-01", "date_label": "November 1966", "lab": "history", "type": "event", "title": "ALPAC report on machine translation", "summary": "A U.S. advisory committee concluded that machine translation had not met expectations, leading to sharp cuts in funding for the work.", "primary_source": "", "secondary_source": "https://en.wikipedia.org/wiki/ALPAC" }, { "date": "1968-01-01", "date_label": "1968", "lab": "history", "type": "literature", "title": "Philip K. Dick's Do Androids Dream of Electric Sheep?", "summary": "Dick's novel follows a bounty hunter tracking androids nearly indistinguishable from humans, raising questions about empathy and identity.", "primary_source": "", "secondary_source": "https://en.wikipedia.org/wiki/Do_Androids_Dream_of_Electric_Sheep%3F" }, { "date": "1968-04-02", "date_label": "2 April 1968", "lab": "history", "type": "film", "title": "2001: A Space Odyssey and HAL 9000", "summary": "Stanley Kubrick's film, written with Arthur C. Clarke, premiered with HAL 9000, a conversational computer that turns against the crew.", "primary_source": "", "secondary_source": "https://en.wikipedia.org/wiki/HAL_9000" }, { "date": "1969-01-01", "date_label": "1969", "lab": "history", "type": "paper", "title": "Minsky and Papert's Perceptrons", "summary": "Marvin Minsky and Seymour Papert analysed the limits of single-layer perceptrons, and the book is often linked to a decline in neural network research.", "primary_source": "", "secondary_source": "https://en.wikipedia.org/wiki/Perceptrons_(book)" }, { "date": "1969-01-01", "date_label": "1969", "lab": "history", "type": "paper", "title": "McCarthy and Hayes describe the frame problem", "summary": "John McCarthy and Patrick Hayes identified the frame problem: how a reasoning system can represent what stays unchanged when an action occurs.", "primary_source": "https://plato.stanford.edu/entries/frame-problem/", "secondary_source": "https://en.wikipedia.org/wiki/Frame_problem" }, { "date": "1970-01-01", "date_label": "1968–1970", "lab": "history", "type": "program", "title": "Winograd's SHRDLU", "summary": "Terry Winograd's SHRDLU understood typed English commands about a simulated world of blocks and could answer questions about it.", "primary_source": "", "secondary_source": "https://en.wikipedia.org/wiki/SHRDLU" }, { "date": "1972-01-01", "date_label": "1972", "lab": "history", "type": "invention", "title": "Prolog programming language", "summary": "Alain Colmerauer and Philippe Roussel created Prolog, a logic programming language widely used in AI and later chosen for Japan's Fifth Generation project.", "primary_source": "", "secondary_source": "https://en.wikipedia.org/wiki/Prolog" }, { "date": "1972-01-01", "date_label": "1972", "lab": "history", "type": "program", "title": "Colby's PARRY", "summary": "Psychiatrist Kenneth Colby built PARRY, a program simulating a person with paranoid schizophrenia; it was later connected to ELIZA for a recorded machine-to-machine conversation.", "primary_source": "https://www.rfc-editor.org/rfc/rfc439", "secondary_source": "https://en.wikipedia.org/wiki/PARRY" }, { "date": "1972-01-01", "date_label": "1970s (1972–1976)", "lab": "history", "type": "program", "title": "MYCIN expert system", "summary": "Stanford's MYCIN used several hundred rules to recommend antibiotics for blood infections, performing comparably to specialists in tests though it was never used clinically.", "primary_source": "", "secondary_source": "https://en.wikipedia.org/wiki/Mycin" }, { "date": "1973-01-01", "date_label": "1973", "lab": "history", "type": "event", "title": "The Lighthill report", "summary": "A review for the UK Science Research Council by James Lighthill judged that AI had failed to deliver on its promises, leading to major funding cuts in Britain.", "primary_source": "http://www.chilton-computing.org.uk/inf/literature/reports/lighthill_report/p001.htm", "secondary_source": "https://en.wikipedia.org/wiki/Lighthill_report" }, { "date": "1974-01-01", "date_label": "c. 1974–1980", "lab": "history", "type": "event", "title": "The first AI winter", "summary": "After the Lighthill report and U.S. funding changes, money and interest in AI research fell sharply for several years.", "primary_source": "", "secondary_source": "https://en.wikipedia.org/wiki/AI_winter" }, { "date": "1980-01-01", "date_label": "1980", "lab": "history", "type": "program", "title": "XCON expert system enters commercial use", "summary": "Digital Equipment Corporation began using XCON (R1), a rule-based system that configured computer orders, a sign of the commercial expert systems boom.", "primary_source": "", "secondary_source": "https://en.wikipedia.org/wiki/Xcon" }, { "date": "1980-01-01", "date_label": "1980", "lab": "history", "type": "paper", "title": "Searle's Chinese room argument", "summary": "Philosopher John Searle argued that a program manipulating symbols by rules would not understand language, even if its answers looked fluent.", "primary_source": "https://plato.stanford.edu/entries/chinese-room/", "secondary_source": "https://en.wikipedia.org/wiki/Chinese_room" }, { "date": "1982-01-01", "date_label": "1982", "lab": "history", "type": "event", "title": "Japan's Fifth Generation Computer project", "summary": "Japan's trade ministry launched a ten-year national programme to build logic-based computers for AI, prompting rival programmes in the U.S. and Europe.", "primary_source": "", "secondary_source": "https://en.wikipedia.org/wiki/Fifth_Generation_Computer_Systems" }, { "date": "1982-04-01", "date_label": "April 1982", "lab": "history", "type": "paper", "title": "Hopfield networks", "summary": "Physicist John Hopfield described a recurrent neural network that stores memories as stable states, helping revive interest in neural networks.", "primary_source": "https://doi.org/10.1073/pnas.79.8.2554", "secondary_source": "https://en.wikipedia.org/wiki/Hopfield_network" }, { "date": "1982-06-25", "date_label": "25 June 1982", "lab": "history", "type": "film", "title": "Blade Runner", "summary": "Ridley Scott's film, adapted from Philip K. Dick's novel, brought artificial humans called replicants to a wide audience.", "primary_source": "", "secondary_source": "https://en.wikipedia.org/wiki/Blade_Runner" }, { "date": "1984-01-01", "date_label": "1984", "lab": "history", "type": "research", "title": "The Cyc project begins", "summary": "Douglas Lenat started Cyc, a long-running effort to hand-encode millions of common-sense facts and rules.", "primary_source": "", "secondary_source": "https://en.wikipedia.org/wiki/Cyc" }, { "date": "1984-07-01", "date_label": "July 1984", "lab": "history", "type": "literature", "title": "William Gibson's Neuromancer", "summary": "Gibson's novel, featuring AIs named Wintermute and Neuromancer, popularised the idea of cyberspace and shaped cyberpunk fiction.", "primary_source": "", "secondary_source": "https://en.wikipedia.org/wiki/Neuromancer" }, { "date": "1984-10-26", "date_label": "26 October 1984", "lab": "history", "type": "film", "title": "The Terminator", "summary": "James Cameron's film about Skynet, a defence AI that turns on humanity, became a common cultural reference for AI risk.", "primary_source": "", "secondary_source": "https://en.wikipedia.org/wiki/The_Terminator" }, { "date": "1986-10-09", "date_label": "9 October 1986", "lab": "history", "type": "paper", "title": "Rumelhart, Hinton and Williams popularise backpropagation", "summary": "A paper in Nature showed how backpropagation of errors could train multi-layer neural networks to learn useful internal representations.", "primary_source": "https://doi.org/10.1038/323533a0", "secondary_source": "https://en.wikipedia.org/wiki/Backpropagation" }, { "date": "1987-01-01", "date_label": "late 1980s – early 1990s", "lab": "history", "type": "event", "title": "The second AI winter", "summary": "The market for specialised Lisp machines collapsed in 1987 and expert systems proved costly to maintain, leading to another drop in AI funding.", "primary_source": "", "secondary_source": "https://en.wikipedia.org/wiki/AI_winter" }, { "date": "1988-01-01", "date_label": "1988", "lab": "history", "type": "program", "title": "Jabberwacky chatbot project begins", "summary": "Rollo Carpenter began Jabberwacky, a chatbot meant to learn conversation from its users; it went online in 1997.", "primary_source": "", "secondary_source": "https://en.wikipedia.org/wiki/Jabberwacky" }, { "date": "1989-01-01", "date_label": "1989", "lab": "history", "type": "paper", "title": "LeCun applies backpropagation to handwritten digits", "summary": "Yann LeCun and colleagues at Bell Labs trained a convolutional neural network to read handwritten ZIP codes.", "primary_source": "https://doi.org/10.1162/neco.1989.1.4.541", "secondary_source": "https://en.wikipedia.org/wiki/Convolutional_neural_network" }, { "date": "1991-11-08", "date_label": "8 November 1991", "lab": "history", "type": "event", "title": "First Loebner Prize contest", "summary": "The first Loebner Prize, an annual Turing-test style competition for chatbots, was held at the Computer Museum in Boston.", "primary_source": "", "secondary_source": "https://en.wikipedia.org/wiki/Loebner_Prize" }, { "date": "1992-01-01", "date_label": "1992", "lab": "history", "type": "program", "title": "Tesauro's TD-Gammon", "summary": "Gerald Tesauro of IBM built TD-Gammon, a neural network that learned backgammon by playing itself and reached near top human level.", "primary_source": "https://doi.org/10.1145/203330.203343", "secondary_source": "https://en.wikipedia.org/wiki/TD-Gammon" }, { "date": "1995-01-01", "date_label": "1995", "lab": "history", "type": "paper", "title": "Support-vector networks", "summary": "Corinna Cortes and Vladimir Vapnik published support-vector machines, a classification method that dominated much of machine learning for the next decade.", "primary_source": "https://doi.org/10.1007/BF00994018", "secondary_source": "https://en.wikipedia.org/wiki/Support_vector_machine" }, { "date": "1995-11-23", "date_label": "23 November 1995", "lab": "history", "type": "program", "title": "Richard Wallace's A.L.I.C.E.", "summary": "Richard Wallace released A.L.I.C.E., a pattern-matching chatbot written in AIML that later won the Loebner Prize three times.", "primary_source": "", "secondary_source": "https://en.wikipedia.org/wiki/Artificial_Linguistic_Internet_Computer_Entity" }, { "date": "1996-05-10", "date_label": "May 1996", "lab": "history", "type": "event", "title": "Brewster Kahle founds the Internet Archive", "summary": "Brewster Kahle started the Internet Archive in San Francisco, a nonprofit library that began crawling and saving the public Web, Netnews and downloadable software before they disappeared.", "primary_source": "https://web.archive.org/web/1997/http://www.archive.org/sciam_article.html", "secondary_source": "https://en.wikipedia.org/wiki/Internet_Archive" }, { "date": "1997-05-11", "date_label": "11 May 1997", "lab": "history", "type": "event", "title": "Deep Blue defeats Garry Kasparov", "summary": "IBM's Deep Blue won a six-game match against world chess champion Garry Kasparov, the first match loss by a reigning champion to a computer under tournament conditions.", "primary_source": "https://www.ibm.com/history/deep-blue", "secondary_source": "https://en.wikipedia.org/wiki/Deep_Blue_versus_Garry_Kasparov" }, { "date": "1997-11-15", "date_label": "15 November 1997", "lab": "history", "type": "paper", "title": "Long short-term memory (LSTM)", "summary": "Sepp Hochreiter and Jürgen Schmidhuber introduced LSTM, a recurrent network design that can retain information over long sequences.", "primary_source": "https://doi.org/10.1162/neco.1997.9.8.1735", "secondary_source": "https://en.wikipedia.org/wiki/Long_short-term_memory" }, { "date": "1998-11-01", "date_label": "November 1998", "lab": "history", "type": "paper", "title": "LeNet-5 and the MNIST dataset", "summary": "LeCun, Bottou, Bengio and Haffner described gradient-based learning for document recognition, including the LeNet-5 network used to read cheques.", "primary_source": "https://doi.org/10.1109/5.726791", "secondary_source": "https://en.wikipedia.org/wiki/LeNet" }, { "date": "1999-08-31", "date_label": "", "lab": "nvidia", "type": "chip", "title": "Nvidia unveils the GeForce 256", "summary": "Nvidia's GeForce 256 graphics processing unit was unveiled, with products from Creative promised for that fall.", "primary_source": "https://sg.creative.com/corporate/pressroom?id=6254", "secondary_source": "https://web.archive.org/web/20241023155821id_/https://money.cnn.com/1999/08/31/technology/nvidia/" }, { "date": "2000-01-01", "date_label": "2000", "lab": "history", "type": "invention", "title": "Honda's ASIMO", "summary": "Honda unveiled ASIMO, a humanoid robot able to walk smoothly and climb stairs.", "primary_source": "", "secondary_source": "https://en.wikipedia.org/wiki/ASIMO" }, { "date": "2000-10-27", "date_label": "", "lab": "ibta", "type": "infrastructure", "title": "InfiniBand 1.0 specification", "summary": "The InfiniBand Trade Association, founded in 1999, released version 1.0 of the InfiniBand Architecture specification.", "primary_source": "https://www.hpcwire.com/2000/10/27/infiniband-trade-assoc-releases-1-0-specification/", "secondary_source": "https://infinibandta.org/about-the-ibta/" }, { "date": "2001-10-24", "date_label": "October 24, 2001", "lab": "history", "type": "invention", "title": "The Wayback Machine opens to the public", "summary": "The Internet Archive unveiled the Wayback Machine at UC Berkeley's Bancroft Library, letting anyone view more than 10 billion archived web pages saved since 1996.", "primary_source": "https://web.archive.org/web/20011201220833/http://www.archive.org:80/wayback/press_kit/press_release.html", "secondary_source": "https://en.wikipedia.org/wiki/Wayback_Machine" }, { "date": "2002-09-01", "date_label": "September 2002", "lab": "history", "type": "invention", "title": "iRobot's Roomba", "summary": "iRobot launched the Roomba vacuum cleaner, which became the first widely sold home robot.", "primary_source": "", "secondary_source": "https://en.wikipedia.org/wiki/Roomba" }, { "date": "2004-03-13", "date_label": "13 March 2004", "lab": "history", "type": "event", "title": "First DARPA Grand Challenge", "summary": "DARPA held a race for driverless vehicles in the Mojave Desert; no vehicle finished, and the best travelled about 12 km.", "primary_source": "", "secondary_source": "https://en.wikipedia.org/wiki/DARPA_Grand_Challenge_(2004)" }, { "date": "2005-10-08", "date_label": "8 October 2005", "lab": "history", "type": "event", "title": "Stanford's Stanley wins the DARPA Grand Challenge", "summary": "Five autonomous vehicles completed the 212 km desert course, with Stanford's Stanley finishing first.", "primary_source": "", "secondary_source": "https://en.wikipedia.org/wiki/DARPA_Grand_Challenge_(2005)" }, { "date": "2006-07-01", "date_label": "July 2006", "lab": "history", "type": "paper", "title": "Hinton's deep belief networks", "summary": "Geoffrey Hinton, Simon Osindero and Yee-Whye Teh showed how to train deep neural networks one layer at a time, a turning point for what became known as deep learning.", "primary_source": "https://doi.org/10.1162/neco.2006.18.7.1527", "secondary_source": "https://en.wikipedia.org/wiki/Deep_belief_network" }, { "date": "2006-11-08", "date_label": "", "lab": "nvidia", "type": "chip", "title": "Nvidia unveils CUDA", "summary": "Nvidia unveiled CUDA, a computing architecture for its GPUs with what it called the industry's first C compiler environment for the GPU.", "primary_source": "https://web.archive.org/web/20080125045805/http://www.nvidia.com:80/object/IO_37226.html", "secondary_source": "" }, { "date": "2007-06-20", "date_label": "", "lab": "nvidia", "type": "chip", "title": "Nvidia launches the Tesla line", "summary": "Nvidia launched Tesla, a line of GPU computing processors aimed at scientific and technical computing rather than graphics.", "primary_source": "https://web.archive.org/web/20090201001351/http://nvidia.com:80/object/IO_43499.html", "secondary_source": "" }, { "date": "2008-06-09", "date_label": "", "lab": "us-doe", "type": "hardware", "title": "Roadrunner passes one petaflop", "summary": "Los Alamos's IBM Roadrunner became the first system to pass one petaflop, reaching 1.026 and taking first place on the TOP500.", "primary_source": "https://top500.org/resources/top-systems/roadrunner-los-alamos-national-laboratory", "secondary_source": "https://www.theregister.com/2008/06/09/roadrunner_supercomputer_debut/" }, { "date": "2008-11-18", "date_label": "", "lab": "llnl", "type": "hardware", "title": "Hyperion testbed announced at Lawrence Livermore", "summary": "Lawrence Livermore and 10 computing companies announced Hyperion, a Linux cluster testbed for next-generation supercomputing.", "primary_source": "https://www.llnl.gov/news/lawrence-livermore-teams-computing-industry-leaders-develop-advanced-technology-cluster-testbed", "secondary_source": "https://hpc.llnl.gov/hardware/platforms/hyperion-decommissioned" }, { "date": "2009-06-01", "date_label": "June 2009", "lab": "history", "type": "research", "title": "ImageNet dataset", "summary": "Fei-Fei Li and colleagues presented ImageNet, a database of millions of labelled images that became the main benchmark for computer vision.", "primary_source": "https://doi.org/10.1109/CVPR.2009.5206848", "secondary_source": "https://en.wikipedia.org/wiki/ImageNet" }, { "date": "2009-09-30", "date_label": "", "lab": "nvidia", "type": "chip", "title": "Nvidia unveils the Fermi architecture", "summary": "Nvidia introduced its next CUDA architecture, codenamed Fermi, built for GPU computing as well as graphics.", "primary_source": "https://web.archive.org/web/20111201204609/http://www.nvidia.com:80/object/io_1254288141829.html", "secondary_source": "" }, { "date": "2010-01-01", "date_label": "2010", "lab": "history", "type": "event", "title": "DeepMind founded", "summary": "Demis Hassabis, Shane Legg and Mustafa Suleyman founded DeepMind in London to research general-purpose learning systems.", "primary_source": "", "secondary_source": "https://en.wikipedia.org/wiki/Google_DeepMind" }, { "date": "2010-10-28", "date_label": "", "lab": "china", "type": "hardware", "title": "China's Tianhe-1A reaches 2.5 petaflops with GPUs", "summary": "China unveiled Tianhe-1A in Tianjin, with 7,168 Nvidia Tesla M2050 GPUs and a sustained 2.5 petaflops.", "primary_source": "https://www.pcworld.com/article/504321/china_unveils_powerful_25petaflop_supercomputer.html", "secondary_source": "https://www.datacenterknowledge.com/business/china-s-tianhe-1a-achieves-2-507-petaflops" }, { "date": "2011-02-16", "date_label": "14–16 February 2011", "lab": "history", "type": "event", "title": "IBM Watson wins Jeopardy!", "summary": "IBM's Watson question-answering system beat champions Ken Jennings and Brad Rutter on the television quiz show Jeopardy!.", "primary_source": "https://www.ibm.com/history/watson-jeopardy", "secondary_source": "https://en.wikipedia.org/wiki/IBM_Watson" }, { "date": "2011-10-04", "date_label": "4 October 2011", "lab": "history", "type": "program", "title": "Siri launches on the iPhone 4S", "summary": "Apple announced Siri, a voice assistant built on technology from a startup it had acquired, as a feature of the iPhone 4S.", "primary_source": "", "secondary_source": "https://en.wikipedia.org/wiki/Siri" }, { "date": "2012-01-01", "date_label": "2012", "lab": "history", "type": "research", "title": "AlexNet wins ImageNet", "summary": "Alex Krizhevsky, Ilya Sutskever and Geoffrey Hinton's convolutional network, trained on GPUs, won the ImageNet challenge by a wide margin, starting the deep learning boom.", "primary_source": "https://papers.nips.cc/paper/2012/hash/c399862d3b9d6b76c8436e924a68c45b-Abstract.html", "secondary_source": "https://en.wikipedia.org/wiki/AlexNet" }, { "date": "2012-06-01", "date_label": "June 2012", "lab": "history", "type": "research", "title": "Google Brain learns to recognise cats", "summary": "A Google and Stanford team trained a large neural network on unlabelled YouTube frames; it learned features responding to faces and cats.", "primary_source": "https://arxiv.org/abs/1112.6209", "secondary_source": "https://en.wikipedia.org/wiki/Google_Brain" }, { "date": "2012-10-29", "date_label": "", "lab": "us-doe", "type": "hardware", "title": "Oak Ridge debuts Titan", "summary": "Oak Ridge debuted Titan, a Cray system with 18,688 nodes, each with an AMD Opteron and an Nvidia Tesla K20 GPU.", "primary_source": "https://www.ornl.gov/news/ornl-debuts-titan-supercomputer", "secondary_source": "https://www.olcf.ornl.gov/2012/10/29/ornl-debuts-titan-supercomputer/" }, { "date": "2012-11-12", "date_label": "", "lab": "nvidia", "type": "chip", "title": "Nvidia Tesla K20 and the Titan supercomputer", "summary": "Nvidia announced the Kepler-based Tesla K20 accelerators, which it said power Titan at Oak Ridge, ranked first in the world.", "primary_source": "https://nvidianews.nvidia.com/news/nvidia-unveils-world-s-fastest-most-efficient-accelerators-powers-world-s-no-1-supercomputer", "secondary_source": "" }, { "date": "2013-01-16", "date_label": "16 January 2013", "lab": "history", "type": "paper", "title": "word2vec", "summary": "Tomas Mikolov and colleagues at Google published word2vec, an efficient way to represent words as vectors that capture meaning from context.", "primary_source": "https://arxiv.org/abs/1301.3781", "secondary_source": "https://en.wikipedia.org/wiki/Word2vec" }, { "date": "2013-10-01", "date_label": "", "lab": "jedec", "type": "chip", "title": "JEDEC publishes the first HBM standard", "summary": "JEDEC published JESD235, the first standard for high-bandwidth memory, DRAM stacked close to the processor over a wide interface.", "primary_source": "https://web.archive.org/web/20131210000000id_/http://www.jedec.org/standards-documents/results/jesd235", "secondary_source": "https://www-old.cs.utah.edu/thememoryforum/mike.pdf" }, { "date": "2013-12-19", "date_label": "19 December 2013", "lab": "history", "type": "paper", "title": "DeepMind's deep Q-network plays Atari", "summary": "DeepMind showed a single neural network learning to play several Atari games directly from screen pixels using reinforcement learning.", "primary_source": "https://arxiv.org/abs/1312.5602", "secondary_source": "https://en.wikipedia.org/wiki/Deep_reinforcement_learning" }, { "date": "2014-01-26", "date_label": "26 January 2014", "lab": "history", "type": "event", "title": "Google acquires DeepMind", "summary": "Google agreed to buy DeepMind, reportedly for around 400 to 650 million U.S. dollars.", "primary_source": "", "secondary_source": "https://en.wikipedia.org/wiki/Google_DeepMind" }, { "date": "2014-03-25", "date_label": "", "lab": "nvidia", "type": "chip", "title": "Nvidia announces NVLink", "summary": "Nvidia announced NVLink, a high-speed GPU interconnect co-developed with IBM, and said it would arrive in the Pascal architecture expected in 2016.", "primary_source": "https://nvidianews.nvidia.com/news/nvidia-launches-world-s-first-high-speed-gpu-interconnect-helping-pave-the-way-to-exascale-computing", "secondary_source": "" }, { "date": "2014-06-07", "date_label": "7 June 2014", "lab": "history", "type": "event", "title": "Eugene Goostman Turing test claim", "summary": "Organisers at the University of Reading claimed the chatbot Eugene Goostman passed a Turing test by convincing a third of judges; many researchers disputed the claim.", "primary_source": "", "secondary_source": "https://en.wikipedia.org/wiki/Eugene_Goostman" }, { "date": "2014-06-10", "date_label": "10 June 2014", "lab": "history", "type": "paper", "title": "Generative adversarial networks", "summary": "Ian Goodfellow and colleagues introduced GANs, in which two networks compete, one generating samples and one judging them, to produce realistic data.", "primary_source": "https://arxiv.org/abs/1406.2661", "secondary_source": "https://en.wikipedia.org/wiki/Generative_adversarial_network" }, { "date": "2014-09-01", "date_label": "1 September 2014", "lab": "history", "type": "paper", "title": "Attention mechanism for neural translation", "summary": "Dzmitry Bahdanau, Kyunghyun Cho and Yoshua Bengio added an attention mechanism letting a translation network focus on relevant source words, a precursor to transformers.", "primary_source": "https://arxiv.org/abs/1409.0473", "secondary_source": "https://en.wikipedia.org/wiki/Attention_(machine_learning)" }, { "date": "2014-09-10", "date_label": "10 September 2014", "lab": "history", "type": "paper", "title": "Sequence-to-sequence learning", "summary": "Ilya Sutskever, Oriol Vinyals and Quoc Le showed that LSTM networks could map one sequence to another, achieving strong machine translation results.", "primary_source": "https://arxiv.org/abs/1409.3215", "secondary_source": "https://en.wikipedia.org/wiki/Seq2seq" }, { "date": "2014-11-06", "date_label": "6 November 2014", "lab": "history", "type": "program", "title": "Amazon Echo and Alexa", "summary": "Amazon introduced the Echo speaker with its voice assistant Alexa, initially by invitation.", "primary_source": "", "secondary_source": "https://en.wikipedia.org/wiki/Amazon_Echo" }, { "date": "2015-06-16", "date_label": "", "lab": "amd", "type": "chip", "title": "AMD Radeon Fury, first GPU family with HBM", "summary": "AMD announced its Radeon 300 series, including what it called the world's first graphics family with High Bandwidth Memory.", "primary_source": "https://ir.amd.com/news-events/press-releases/detail/619/amd-ushers-in-a-new-era-of-pc-gaming-with-radeontm-r9-and-r7-300-series-graphics-line-up-including-worlds-first-graphics-family-with-revolutionary-hbm-technology", "secondary_source": "" }, { "date": "2015-11-09", "date_label": "9 November 2015", "lab": "history", "type": "invention", "title": "Google open-sources TensorFlow", "summary": "Google released TensorFlow, its machine learning software library, as open source.", "primary_source": "", "secondary_source": "https://en.wikipedia.org/wiki/TensorFlow" }, { "date": "2015-11-16", "date_label": "", "lab": "amd", "type": "software", "title": "AMD launches the Boltzmann Initiative", "summary": "AMD announced a compiler, a headless Linux driver and the HIP tool for porting CUDA code, aimed at GPU computing on its FirePro cards.", "primary_source": "https://ir.amd.com/news-events/press-releases/detail/652/amd-launches-boltzmann-initiative-to-dramatically-reduce-barriers-to-gpu-computing-on-amd-fireprotm-graphics", "secondary_source": "https://www.guru3d.com/story/amd-launches-boltzmann-initiative-hsa-porting-cuda-based-applications/" }, { "date": "2015-12-10", "date_label": "10 December 2015", "lab": "history", "type": "paper", "title": "Residual networks (ResNet)", "summary": "Kaiming He and colleagues at Microsoft Research introduced residual connections that made very deep networks trainable, winning ImageNet 2015.", "primary_source": "https://arxiv.org/abs/1512.03385", "secondary_source": "https://en.wikipedia.org/wiki/Residual_neural_network" }, { "date": "2015-12-11", "date_label": "", "lab": "openai", "type": "event", "title": "OpenAI founded", "summary": "OpenAI was announced as a non-profit AI research lab with backing from several tech investors.", "primary_source": "https://openai.com/index/introducing-openai/", "secondary_source": "https://en.wikipedia.org/wiki/OpenAI" }, { "date": "2016-01-12", "date_label": "", "lab": "jedec", "type": "chip", "title": "JEDEC updates the HBM standard (HBM2)", "summary": "JEDEC published JESD235A, an update to the High Bandwidth Memory standard, supporting up to 8 GB per device at up to 256 GB per second.", "primary_source": "https://www.jedec.org/news/pressreleases/jedec-updates-groundbreaking-high-bandwidth-memory-hbm-standard", "secondary_source": "" }, { "date": "2016-01-19", "date_label": "", "lab": "samsung", "type": "chip", "title": "Samsung starts mass-producing HBM2", "summary": "Samsung said it had begun mass-producing the industry's first 4 GB DRAM package using the HBM2 interface.", "primary_source": "https://news.samsung.com/global/samsung-begins-mass-producing-worlds-fastest-dram-based-on-newest-high-bandwidth-memory-hbm-interface", "secondary_source": "https://www.storagereview.com/news/samsung-begins-mass-production-of-hbm2-4gb-dram-package" }, { "date": "2016-01-27", "date_label": "27 January 2016", "lab": "history", "type": "paper", "title": "AlphaGo paper and win over Fan Hui", "summary": "DeepMind published AlphaGo in Nature, revealing it had beaten European champion Fan Hui 5–0 in October 2015, the first win by a program over a professional Go player on a full board without handicap.", "primary_source": "https://doi.org/10.1038/nature16961", "secondary_source": "https://en.wikipedia.org/wiki/AlphaGo_versus_Fan_Hui" }, { "date": "2016-03-15", "date_label": "9–15 March 2016", "lab": "history", "type": "event", "title": "AlphaGo defeats Lee Sedol", "summary": "DeepMind's AlphaGo beat Go champion Lee Sedol four games to one in Seoul, a milestone many experts had expected to be a decade away.", "primary_source": "", "secondary_source": "https://en.wikipedia.org/wiki/AlphaGo_versus_Lee_Sedol" }, { "date": "2016-03-23", "date_label": "23 March 2016", "lab": "history", "type": "program", "title": "Microsoft's Tay chatbot", "summary": "Microsoft launched Tay, a Twitter chatbot that learned from users; it was taken offline within a day after being manipulated into posting offensive messages.", "primary_source": "", "secondary_source": "https://en.wikipedia.org/wiki/Tay_(chatbot)" }, { "date": "2016-04-05", "date_label": "", "lab": "nvidia", "type": "hardware", "title": "Nvidia DGX-1 and Tesla P100 (Pascal)", "summary": "Nvidia launched the DGX-1 deep learning system, built on Pascal-based Tesla P100 GPUs.", "primary_source": "https://nvidianews.nvidia.com/news/nvidia-launches-world-s-first-deep-learning-supercomputer", "secondary_source": "" }, { "date": "2016-05-18", "date_label": "", "lab": "google", "type": "chip", "title": "Google reveals the TPU", "summary": "Google revealed the Tensor Processing Unit, a custom chip for machine learning that it said was already running in its data centres.", "primary_source": "https://cloud.google.com/blog/products/ai-machine-learning/google-supercharges-machine-learning-tasks-with-custom-chip", "secondary_source": "" }, { "date": "2016-06-20", "date_label": "", "lab": "china", "type": "hardware", "title": "Sunway TaihuLight tops the TOP500", "summary": "China's Sunway TaihuLight took first place at 93 petaflops, using processors designed and made in China.", "primary_source": "https://www.top500.org/news/new-chinese-supercomputer-named-worlds-fastest-system-on-latest-top500-list/", "secondary_source": "https://top500.org/lists/top500/2016/06/highlights/" }, { "date": "2016-08-09", "date_label": "", "lab": "intel", "type": "finance", "title": "Intel buys Nervana Systems", "summary": "Intel announced it was buying the deep learning startup Nervana, which was designing a custom training chip; press reports valued the deal above $350 million.", "primary_source": "https://top500.org/news/intel-buys-deep-learning-specialist-nervana-to-boost-ai-portfolio/", "secondary_source": "https://techcrunch.com/2016/08/09/intel-buys-deep-learning-startup-nervana-systems-for-a-reported-350-million" }, { "date": "2016-09-08", "date_label": "8 September 2016", "lab": "history", "type": "paper", "title": "DeepMind's WaveNet", "summary": "DeepMind introduced WaveNet, a neural network that generates raw audio waveforms and produced more natural synthetic speech.", "primary_source": "https://arxiv.org/abs/1609.03499", "secondary_source": "https://en.wikipedia.org/wiki/WaveNet" }, { "date": "2016-09-27", "date_label": "27 September 2016", "lab": "history", "type": "research", "title": "Google Neural Machine Translation", "summary": "Google announced a neural machine translation system, first used for Chinese to English in Google Translate, replacing its phrase-based approach.", "primary_source": "https://arxiv.org/abs/1609.08144", "secondary_source": "https://en.wikipedia.org/wiki/Google_Neural_Machine_Translation" }, { "date": "2016-11-14", "date_label": "", "lab": "amd", "type": "software", "title": "AMD releases ROCm 1.3 at SC16", "summary": "AMD released a new version of its open-source Radeon Open Compute platform, adding Polaris GPU support, Docker containers and a CUDA porting path through HIP.", "primary_source": "https://ir.amd.com/news-events/press-releases/detail/732/amd-releases-new-version-of-rocm-the-most-versatile-open-source-platform-for-gpu-computing", "secondary_source": "https://semiaccurate.com/2016/11/14/amd-rocm-botlzmann-v1-3/" }, { "date": "2016-12-12", "date_label": "", "lab": "amd", "type": "chip", "title": "AMD Radeon Instinct MI6, MI8 and MI25", "summary": "AMD introduced Radeon Instinct, its accelerator brand for machine intelligence, with the MI6, MI8 and MI25 and the MIOpen library.", "primary_source": "https://ir.amd.com/news-events/press-releases/detail/740/amd-introduces-radeon-instinct-accelerating-machine-intelligence", "secondary_source": "" }, { "date": "2017-04-16", "date_label": "", "lab": "google", "type": "chip", "title": "Google publishes TPU performance paper", "summary": "Google researchers published a paper analysing the performance of the first TPU in its data centres.", "primary_source": "https://arxiv.org/abs/1704.04760", "secondary_source": "" }, { "date": "2017-05-10", "date_label": "", "lab": "nvidia", "type": "chip", "title": "Nvidia launches Volta and the Tesla V100", "summary": "Nvidia launched the Volta architecture and the Tesla V100 data centre GPU, which it rated above 120 teraflops of deep learning performance.", "primary_source": "https://nvidianews.nvidia.com/news/nvidia-launches-revolutionary-volta-gpu-platform-fueling-next-era-of-ai-and-high-performance-computing", "secondary_source": "" }, { "date": "2017-05-17", "date_label": "", "lab": "google", "type": "chip", "title": "Second-generation Cloud TPU", "summary": "Google announced second-generation Cloud TPUs, each rated at up to 180 teraflops, for both training and inference.", "primary_source": "https://blog.google/innovation-and-ai/infrastructure-and-cloud/google-cloud/google-cloud-offer-tpus-machine-learning/", "secondary_source": "" }, { "date": "2017-06-12", "date_label": "", "lab": "google", "type": "paper", "title": "Attention Is All You Need", "summary": "Google researchers introduced the Transformer architecture, which underlies modern large language models.", "primary_source": "https://arxiv.org/abs/1706.03762", "secondary_source": "https://en.wikipedia.org/wiki/Attention_Is_All_You_Need" }, { "date": "2017-09-12", "date_label": "", "lab": "apple", "type": "chip", "title": "Apple A11 Bionic with Neural Engine", "summary": "Apple announced the iPhone X with the A11 Bionic chip and its Neural Engine.", "primary_source": "https://www.apple.com/newsroom/2017/09/the-future-is-here-iphone-x/", "secondary_source": "" }, { "date": "2017-12-07", "date_label": "", "lab": "nvidia", "type": "hardware", "title": "Nvidia Titan V", "summary": "Nvidia introduced the Titan V, a Volta-based GPU for desktop PCs aimed at AI researchers.", "primary_source": "https://nvidianews.nvidia.com/news/nvidia-titan-v-transforms-the-pc-into-ai-supercomputer", "secondary_source": "" }, { "date": "2018-01-01", "date_label": "", "lab": "tsmc", "type": "chip", "title": "TSMC starts 7 nm volume production", "summary": "TSMC says it was the first foundry to start 7-nanometre FinFET volume production, in 2018.", "primary_source": "https://www.tsmc.com/english/dedicatedFoundry/technology/logic/l_7nm", "secondary_source": "" }, { "date": "2018-02-12", "date_label": "", "lab": "google", "type": "product", "title": "Cloud TPUs open in beta", "summary": "Google made Cloud TPUs available in beta on Google Cloud, so outside customers could rent its accelerators.", "primary_source": "https://cloud.google.com/blog/products/gcp/cloud-tpu-machine-learning-accelerators-now-available-in-beta", "secondary_source": "" }, { "date": "2018-03-27", "date_label": "", "lab": "nvidia", "type": "hardware", "title": "Nvidia DGX-2 and NVSwitch", "summary": "At GTC, Nvidia unveiled the DGX-2, a system joining 16 GPUs through its new NVSwitch fabric, which it rated at 2 petaflops.", "primary_source": "https://nvidianews.nvidia.com/news/nvidia-boosts-worlds-leading-deep-learning-computing-platform-bringing-10x-performance-gain-in-six-months", "secondary_source": "" }, { "date": "2018-05-08", "date_label": "", "lab": "google", "type": "chip", "title": "Google TPU 3.0", "summary": "At Google I/O, Sundar Pichai announced the third-generation TPU, with pods he said were eight times more powerful, and liquid cooling in Google's data centres.", "primary_source": "https://techcrunch.com/2018/05/08/google-announces-a-new-generation-for-its-tpu-machine-learning-hardware/", "secondary_source": "https://top500.org/news/google-offers-glimpse-of-third-generation-tpu-processor/" }, { "date": "2018-05-29", "date_label": "", "lab": "nvidia", "type": "hardware", "title": "Nvidia HGX-2 platform", "summary": "Nvidia introduced HGX-2, a platform for up to 16 GPUs; the DGX-2 was the first system built on it.", "primary_source": "https://nvidianews.nvidia.com/news/nvidia-introduces-hgx-2-fusing-hpc-and-ai-computing-into-unified-architecture-6696445", "secondary_source": "" }, { "date": "2018-06-08", "date_label": "", "lab": "us-doe", "type": "infrastructure", "title": "Summit supercomputer launched", "summary": "Oak Ridge launched Summit, with 27,648 Nvidia V100 GPUs and a peak of 200 petaflops.", "primary_source": "https://www.ornl.gov/news/ornl-launches-summit-supercomputer", "secondary_source": "" }, { "date": "2018-06-11", "date_label": "", "lab": "openai", "type": "paper", "title": "GPT-1: generative pre-training paper", "summary": "OpenAI described a transformer language model pre-trained on unlabeled text and then fine-tuned for specific tasks, the first model later called GPT.", "primary_source": "https://openai.com/index/language-unsupervised/", "secondary_source": "https://en.wikipedia.org/wiki/GPT-1" }, { "date": "2018-08-20", "date_label": "", "lab": "nvidia", "type": "chip", "title": "GeForce RTX 20 series (Turing)", "summary": "Nvidia announced the GeForce RTX 2080 Ti, 2080 and 2070, its first consumer RTX cards.", "primary_source": "https://nvidianews.nvidia.com/news/10-years-in-the-making-nvidia-brings-real-time-ray-tracing-to-gamers-with-geforce-rtx", "secondary_source": "" }, { "date": "2018-10-11", "date_label": "", "lab": "google", "type": "paper", "title": "BERT paper", "summary": "Google introduced BERT, a bidirectional transformer pre-trained on text, which set records on language understanding tasks.", "primary_source": "https://arxiv.org/abs/1810.04805", "secondary_source": "https://en.wikipedia.org/wiki/BERT_(language_model)" }, { "date": "2018-11-06", "date_label": "", "lab": "amd", "type": "chip", "title": "AMD Radeon Instinct MI60 and MI50", "summary": "AMD unveiled the MI60 and MI50, which it called the first 7-nanometre data centre GPUs.", "primary_source": "https://ir.amd.com/news-events/press-releases/detail/859/amd-unveils-worlds-first-7nm-datacenter-gpus----powering-the-next-era-of-artificial-intelligence-cloud-computing-and-high-performance-computing-hpc", "secondary_source": "" }, { "date": "2018-11-28", "date_label": "", "lab": "amazon", "type": "chip", "title": "AWS announces Inferentia", "summary": "AWS announced Inferentia, a custom machine learning inference chip.", "primary_source": "https://aws.amazon.com/about-aws/whats-new/2018/11/announcing-amazon-inferentia-machine-learning-inference-microchip/", "secondary_source": "" }, { "date": "2019-02-14", "date_label": "", "lab": "openai", "type": "model", "title": "GPT-2 announced with staged release", "summary": "OpenAI unveiled the 1.5-billion-parameter GPT-2 but released only a small version, citing misuse concerns.", "primary_source": "https://openai.com/index/better-language-models/", "secondary_source": "https://en.wikipedia.org/wiki/GPT-2" }, { "date": "2019-02-19", "date_label": "", "lab": "jedec", "type": "chip", "title": "JEDEC publishes LPDDR5", "summary": "JEDEC published JESD209-5, a low-power memory standard targeting 6,400 MT/s.", "primary_source": "https://www.jedec.org/news/pressreleases/jedec-updates-standard-low-power-memory-devices-lpddr5", "secondary_source": "" }, { "date": "2019-03-06", "date_label": "", "lab": "google", "type": "hardware", "title": "Google opens Coral, with the Edge TPU, in public beta", "summary": "Google put Coral in public beta, offering local AI hardware built on its small Edge TPU chip.", "primary_source": "https://developers.googleblog.com/2019/03/introducing-coral-our-platform-for.html", "secondary_source": "https://insidehpc.com/2019/03/announcing-googles-new-tpu-dev-board-for-machine-learning-on-the-edge" }, { "date": "2019-03-11", "date_label": "", "lab": "openai", "type": "policy", "title": "OpenAI LP capped-profit structure", "summary": "OpenAI created a capped-profit company under its non-profit to raise outside capital.", "primary_source": "https://openai.com/index/openai-lp/", "secondary_source": "https://en.wikipedia.org/wiki/OpenAI" }, { "date": "2019-03-11", "date_label": "", "lab": "nvidia", "type": "infrastructure", "title": "Nvidia agrees to buy Mellanox", "summary": "Nvidia agreed to acquire Mellanox, a maker of InfiniBand and Ethernet data centre networking, for 6.9 billion US dollars.", "primary_source": "https://nvidianews.nvidia.com/news/nvidia-to-acquire-mellanox-for-6-9-billion", "secondary_source": "" }, { "date": "2019-03-18", "date_label": "", "lab": "nvidia", "type": "hardware", "title": "Nvidia announces the $99 Jetson Nano", "summary": "Nvidia announced a small $99 AI computer that delivers 472 GFLOPS.", "primary_source": "https://nvidianews.nvidia.com/news/nvidia-announces-jetson-nano-99-tiny-yet-mighty-nvidia-cuda-x-ai-computer-that-runs-all-ai-models", "secondary_source": "https://www.theregister.com/2019/03/19/nvidia_gtc_news/" }, { "date": "2019-04-25", "date_label": "", "lab": "openai", "type": "model", "title": "MuseNet music model", "summary": "OpenAI showed a transformer that composes multi-instrument music in several styles.", "primary_source": "https://openai.com/index/musenet/", "secondary_source": "https://en.wikipedia.org/wiki/OpenAI" }, { "date": "2019-05-16", "date_label": "", "lab": "us-commerce", "type": "policy", "title": "US adds Huawei to the Entity List", "summary": "The US Commerce Department added Huawei and affiliates to the Entity List, requiring licences for US technology exports to them.", "primary_source": "https://www.federalregister.gov/documents/2019/05/21/2019-10616/addition-of-entities-to-the-entity-list", "secondary_source": "" }, { "date": "2019-06-17", "date_label": "", "lab": "nvidia", "type": "hardware", "title": "Nvidia DGX SuperPOD reaches 22nd on the TOP500", "summary": "Nvidia described a cluster of 96 DGX-2H servers and 1,536 V100 GPUs that scored 9.444 petaflops and placed 22nd on the June 2019 list.", "primary_source": "https://developer.nvidia.com/blog/dgx-superpod-world-record-supercomputing-enterprise/", "secondary_source": "https://www.embedded-computing.com/news/nvidia-dgx-superpod-world-s-22nd-fastest-supercomputer-to-support-company-s-autonomous-vehicle-program" }, { "date": "2019-07-01", "date_label": "", "lab": "japan", "type": "policy", "title": "Japan tightens chip-material exports to South Korea", "summary": "Japan's trade ministry moved fluorinated polyimide, resists and hydrogen fluoride bound for South Korea to individual export licences.", "primary_source": "https://www2.jiia.or.jp/en/article_page.php?id=25", "secondary_source": "https://www.meti.go.jp/english/press/2023/0323_001.html" }, { "date": "2019-07-22", "date_label": "", "lab": "openai", "type": "event", "title": "Microsoft invests $1 billion in OpenAI", "summary": "Microsoft agreed to invest one billion dollars and become OpenAI's exclusive cloud provider.", "primary_source": "https://openai.com/index/microsoft-invests-in-and-partners-with-openai/", "secondary_source": "https://en.wikipedia.org/wiki/OpenAI" }, { "date": "2019-08-19", "date_label": "", "lab": "cerebras", "type": "chip", "title": "Cerebras Wafer Scale Engine", "summary": "Cerebras unveiled the Wafer Scale Engine, a single chip with more than 1.2 trillion transistors and an area of 46,225 square millimetres.", "primary_source": "https://www.cerebras.ai/press-release/cerebras-systems-unveils-the-industrys-first-trillion-transistor-chip", "secondary_source": "" }, { "date": "2019-08-20", "date_label": "", "lab": "openai", "type": "model", "title": "GPT-2 774M model released", "summary": "As part of the staged release, OpenAI published the 774-million-parameter GPT-2 model with a six-month report.", "primary_source": "https://openai.com/index/gpt-2-6-month-follow-up/", "secondary_source": "https://en.wikipedia.org/wiki/GPT-2" }, { "date": "2019-08-23", "date_label": "", "lab": "huawei", "type": "chip", "title": "Huawei Ascend 910", "summary": "Huawei launched the Ascend 910 AI processor, rated at 256 teraflops, and the MindSpore framework.", "primary_source": "https://www.mindspore.cn/news/en/172", "secondary_source": "" }, { "date": "2019-09-25", "date_label": "", "lab": "alibaba", "type": "chip", "title": "Alibaba unveils the Hanguang 800", "summary": "Alibaba's chip unit Pingtouge unveiled its first AI inference chip at the Apsara Conference, for use in Alibaba's cloud.", "primary_source": "https://www.alibabacloud.com/blog/announcing-hanguang-800-alibabas-first-ai-inference-chip_595482", "secondary_source": "https://syncedreview.com/2019/09/25/alibabas-new-ai-chip-can-process-nearly-80k-images-per-second/" }, { "date": "2019-10-25", "date_label": "", "lab": "google", "type": "product", "title": "BERT used in Google Search", "summary": "Google said it was using BERT to better understand search queries.", "primary_source": "https://blog.google/products/search/search-language-understanding-bert/", "secondary_source": "https://en.wikipedia.org/wiki/BERT_(language_model)" }, { "date": "2019-11-05", "date_label": "", "lab": "openai", "type": "model", "title": "Full GPT-2 (1.5B) released", "summary": "OpenAI completed the staged release by publishing the full 1.5-billion-parameter GPT-2 weights.", "primary_source": "https://openai.com/index/gpt-2-1-5b-release/", "secondary_source": "https://en.wikipedia.org/wiki/GPT-2" }, { "date": "2019-12-03", "date_label": "", "lab": "amazon", "type": "chip", "title": "AWS Inf1 instances", "summary": "AWS launched EC2 Inf1 instances, the first cloud servers using its Inferentia chips.", "primary_source": "https://aws.amazon.com/blogs/aws/amazon-ec2-update-inf1-instances-with-aws-inferentia-chips-for-high-performance-cost-effective-inferencing/", "secondary_source": "" }, { "date": "2019-12-16", "date_label": "", "lab": "intel", "type": "chip", "title": "Intel buys Habana Labs", "summary": "Intel bought Israeli AI chipmaker Habana Labs for about 2 billion US dollars.", "primary_source": "https://www.fiercesensors.com/electronics/intel-buys-habana-for-ai-but-what-happens-to-nervana-chips", "secondary_source": "https://www.nextplatform.com/ai/2019/12/17/intel-hedges-its-ai-bets-with-habana-labs-buy/1632965" }, { "date": "2020-01-01", "date_label": "", "lab": "tsmc", "type": "chip", "title": "TSMC starts 5 nm volume production", "summary": "TSMC says it was the first foundry to move 5-nanometre FinFET into volume production, in 2020.", "primary_source": "https://www.tsmc.com/english/dedicatedFoundry/technology/logic/l_5nm", "secondary_source": "" }, { "date": "2020-04-27", "date_label": "", "lab": "nvidia", "type": "infrastructure", "title": "Nvidia completes Mellanox purchase", "summary": "Nvidia completed its acquisition of Mellanox.", "primary_source": "https://nvidianews.nvidia.com/news/nvidia-completes-acquisition-of-mellanox-creating-major-force-driving-next-gen-data-centers", "secondary_source": "" }, { "date": "2020-04-30", "date_label": "", "lab": "openai", "type": "model", "title": "Jukebox music generation model", "summary": "OpenAI released a model that generates raw audio songs, including rudimentary singing.", "primary_source": "https://openai.com/index/jukebox/", "secondary_source": "https://en.wikipedia.org/wiki/OpenAI" }, { "date": "2020-05-14", "date_label": "", "lab": "nvidia", "type": "chip", "title": "Nvidia Ampere and the A100", "summary": "Nvidia announced the Ampere architecture and the A100 data centre GPU, with more than 54 billion transistors on a 7-nanometre process, in full production.", "primary_source": "https://nvidianews.nvidia.com/news/nvidias-new-ampere-data-center-gpu-in-full-production", "secondary_source": "" }, { "date": "2020-05-15", "date_label": "", "lab": "tsmc", "type": "infrastructure", "title": "TSMC announces Arizona fab", "summary": "TSMC said it intended to build a 5-nanometre chip plant in Arizona.", "primary_source": "https://pr.tsmc.com/english/news/2033", "secondary_source": "" }, { "date": "2020-05-15", "date_label": "", "lab": "us-commerce", "type": "policy", "title": "US foreign direct product rule on Huawei", "summary": "The US amended its export rules so that some chips made abroad with US technology for Huawei designs need a licence.", "primary_source": "https://www.federalregister.gov/documents/2020/05/19/2020-10856/export-administration-regulations-amendments-to-general-prohibition-three-foreign-produced-direct", "secondary_source": "" }, { "date": "2020-05-19", "date_label": "", "lab": "microsoft", "type": "infrastructure", "title": "Microsoft's supercomputer for OpenAI", "summary": "Microsoft announced a supercomputer built with OpenAI, with more than 285,000 CPU cores and 10,000 GPUs.", "primary_source": "https://news.microsoft.com/source/features/ai/openai-azure-supercomputer/", "secondary_source": "" }, { "date": "2020-05-28", "date_label": "", "lab": "openai", "type": "paper", "title": "GPT-3 paper: Language Models are Few-Shot Learners", "summary": "OpenAI described the 175-billion-parameter GPT-3 and showed it could do tasks from a few examples in the prompt.", "primary_source": "https://arxiv.org/abs/2005.14165", "secondary_source": "https://en.wikipedia.org/wiki/GPT-3" }, { "date": "2020-06-01", "date_label": "", "lab": "nvidia", "type": "hardware", "title": "Nvidia's Selene debuts at seventh on the TOP500", "summary": "Selene, built from DGX A100 systems with A100 GPUs, entered the June 2020 list at seventh with 27.58 petaflops.", "primary_source": "https://top500.org/system/179842", "secondary_source": "https://www.hpcwire.com/2020/06/22/nvidia-nabs-7-spot-on-top500-with-selene-launches-a100-pcie-cards/" }, { "date": "2020-06-11", "date_label": "", "lab": "openai", "type": "api", "title": "OpenAI API launched (GPT-3 private beta)", "summary": "OpenAI opened a commercial API giving waitlisted developers access to GPT-3 models.", "primary_source": "https://openai.com/index/openai-api/", "secondary_source": "https://en.wikipedia.org/wiki/GPT-3" }, { "date": "2020-06-23", "date_label": "", "lab": "riken", "type": "infrastructure", "title": "Fugaku takes first place on the TOP500", "summary": "RIKEN said Fugaku, built with Fujitsu on Arm technology, had taken first place on the TOP500 with 415.53 petaflops.", "primary_source": "https://www.riken.jp/en/news_pubs/news/2020/20200623_1/index.html", "secondary_source": "" }, { "date": "2020-07-02", "date_label": "", "lab": "sk-hynix", "type": "chip", "title": "SK hynix starts HBM2E mass production", "summary": "SK hynix began mass production of HBM2E, at 16 GB per stack and over 460 GB/s.", "primary_source": "https://news.skhynix.com/sk-hynix-starts-mass-production-of-high-speed-dram-hbm2e/", "secondary_source": "" }, { "date": "2020-07-14", "date_label": "", "lab": "jedec", "type": "chip", "title": "JEDEC publishes DDR5", "summary": "JEDEC published JESD79-5, the DDR5 standard, promising twice the performance of the previous generation for servers and client systems.", "primary_source": "https://www.jedec.org/news/pressreleases/jedec-publishes-new-ddr5-standard-advancing-next-generation-high-performance", "secondary_source": "" }, { "date": "2020-09-01", "date_label": "", "lab": "nvidia", "type": "chip", "title": "GeForce RTX 30 series", "summary": "Nvidia announced the GeForce RTX 3090, 3080 and 3070, built on Ampere.", "primary_source": "https://nvidianews.nvidia.com/news/nvidia-delivers-greatest-ever-generational-leap-in-performance-with-geforce-rtx-30-series-gpus", "secondary_source": "" }, { "date": "2020-09-13", "date_label": "", "lab": "nvidia", "type": "finance", "title": "Nvidia agrees to buy Arm for $40 billion", "summary": "Nvidia announced a deal to buy the chip designer Arm from SoftBank for $40 billion.", "primary_source": "https://nvidianews.nvidia.com/news/nvidia-to-acquire-arm-for-40-billion-creating-worlds-premier-computing-company-for-the-age-of-ai", "secondary_source": "https://embeddedcomputing.com/technology/processing/nvidia-and-softbank-group-announce-termination-of-nvidias-acquisition-of-arm-limited" }, { "date": "2020-10-05", "date_label": "", "lab": "nvidia", "type": "chip", "title": "Nvidia launches the BlueField-2 DPU family", "summary": "Nvidia announced data processing units and the DOCA software kit, moving networking, storage and security work off the CPU.", "primary_source": "https://nvidianews.nvidia.com/news/nvidia-introduces-new-family-of-bluefield-dpus-to-bring-breakthrough-networking-storage-and-security-performance-to-every-data-center", "secondary_source": "" }, { "date": "2020-11-10", "date_label": "", "lab": "apple", "type": "chip", "title": "Apple M1 and unified memory", "summary": "Apple unveiled the M1, its first Mac chip, with a unified memory architecture and a 16-core Neural Engine rated at 11 trillion operations per second.", "primary_source": "https://www.apple.com/newsroom/2020/11/apple-unleashes-m1/", "secondary_source": "" }, { "date": "2020-11-16", "date_label": "", "lab": "nvidia", "type": "infrastructure", "title": "Nvidia introduces 400G InfiniBand", "summary": "Nvidia introduced its next generation of Mellanox InfiniBand at 400 Gb/s at SC20.", "primary_source": "https://nvidianews.nvidia.com/news/nvidia-announces-mellanox-infiniband-for-exascale-ai-supercomputing", "secondary_source": "https://www.servethehome.com/nvidia-mellanox-ndr-400gbps-infiniband-announced/" }, { "date": "2020-11-16", "date_label": "", "lab": "amd", "type": "chip", "title": "AMD Instinct MI100", "summary": "AMD announced the Instinct MI100 accelerator on its CDNA architecture, claiming more than 10 teraflops of double-precision performance.", "primary_source": "https://ir.amd.com/news-events/press-releases/detail/981/amd-announces-worlds-fastest-hpc-accelerator-for-scientific-research", "secondary_source": "" }, { "date": "2020-12-01", "date_label": "", "lab": "amazon", "type": "chip", "title": "AWS announces Trainium", "summary": "At re:Invent, AWS announced Trainium, its custom chip for machine learning training.", "primary_source": "https://techcrunch.com/2020/12/01/aws-launches-trainium-its-new-custom-ml-training-chip/", "secondary_source": "https://www.bio-itworld.com/news/2020/12/14/new-ai-chips-managed-services-among-flood-from-aws-at-re-invent-2020" }, { "date": "2020-12-18", "date_label": "", "lab": "us-commerce", "type": "policy", "title": "US adds SMIC to the Entity List", "summary": "The US added China's top chipmaker, SMIC, and other entities to the Entity List, citing China's military-civil fusion doctrine.", "primary_source": "https://www.govinfo.gov/content/pkg/FR-2020-12-22/html/2020-28031.htm", "secondary_source": "https://www.federalregister.gov/api/v1/documents/2020-28031.json" }, { "date": "2021-01-05", "date_label": "", "lab": "openai", "type": "model", "title": "CLIP released", "summary": "OpenAI released CLIP, a model that learns to link images and text from internet data.", "primary_source": "https://openai.com/index/clip/", "secondary_source": "https://en.wikipedia.org/wiki/Contrastive_Language-Image_Pre-training" }, { "date": "2021-01-05", "date_label": "", "lab": "openai", "type": "model", "title": "DALL-E announced", "summary": "OpenAI showed a GPT-3-style model that creates images from text descriptions.", "primary_source": "https://openai.com/index/dall-e/", "secondary_source": "https://en.wikipedia.org/wiki/DALL-E" }, { "date": "2021-04-01", "date_label": "", "lab": "cerebras", "type": "chip", "title": "Cerebras WSE-2", "summary": "Cerebras unveiled the WSE-2, with 2.6 trillion transistors and 850,000 cores.", "primary_source": "https://www.cerebras.ai/press-release/cerebras-systems-smashes-the-2-5-trillion-transistor-mark-with-new-second-generation-wafer-scale-engine", "secondary_source": "" }, { "date": "2021-04-12", "date_label": "", "lab": "nvidia", "type": "chip", "title": "Nvidia announces the Grace CPU", "summary": "Nvidia announced Grace, its first data centre CPU, for giant AI and high-performance computing workloads.", "primary_source": "https://nvidianews.nvidia.com/news/nvidia-announces-cpu-for-giant-ai-and-high-performance-computing-workloads", "secondary_source": "" }, { "date": "2021-05-18", "date_label": "", "lab": "google", "type": "model", "title": "LaMDA unveiled", "summary": "Google showed LaMDA, a conversational language model, at Google I/O.", "primary_source": "https://blog.google/technology/ai/lamda/", "secondary_source": "https://en.wikipedia.org/wiki/LaMDA" }, { "date": "2021-05-18", "date_label": "", "lab": "google", "type": "chip", "title": "Google TPU v4", "summary": "At Google I/O, Sundar Pichai announced TPU v4, saying one pod delivers more than one exaflop.", "primary_source": "https://blog.google/innovation-and-ai/technology/developers-tools/io21-helpful-google/", "secondary_source": "" }, { "date": "2021-05-27", "date_label": "", "lab": "us-doe", "type": "hardware", "title": "Berkeley Lab dedicates Perlmutter", "summary": "NERSC dedicated Perlmutter, an HPE Cray system whose first phase has 1,536 nodes, each with four Nvidia A100 GPUs.", "primary_source": "https://nersc.gov/news-and-events/news/berkeley-lab-deploys-next-generation-supercomputer-perlmutter-bolstering-u-s-scientific-research", "secondary_source": "https://elementsarchive.lbl.gov/2021/05/27/perlmutter-supercomputer-dedication-today/" }, { "date": "2021-07-07", "date_label": "", "lab": "openai", "type": "paper", "title": "Codex paper: Evaluating LLMs Trained on Code", "summary": "OpenAI published research on Codex, a GPT model fine-tuned on public code, and introduced the HumanEval benchmark.", "primary_source": "https://arxiv.org/abs/2107.03374", "secondary_source": "https://en.wikipedia.org/wiki/OpenAI_Codex" }, { "date": "2021-08-10", "date_label": "", "lab": "openai", "type": "api", "title": "OpenAI Codex in the API", "summary": "OpenAI released an improved Codex model to API beta users, able to turn natural language into code.", "primary_source": "https://openai.com/index/openai-codex/", "secondary_source": "https://en.wikipedia.org/wiki/OpenAI_Codex" }, { "date": "2021-08-19", "date_label": "", "lab": "tesla", "type": "chip", "title": "Tesla reveals the D1 chip for Dojo", "summary": "At AI Day, Tesla showed its D1 training chip, a 7 nm design rated at 362 teraflops, with 25 chips forming one training tile.", "primary_source": "https://www.cnbc.com/2021/08/19/tesla-unveils-dojo-d1-chip-at-ai-day.html", "secondary_source": "https://www.infoq.com/news/2021/09/tesla-dojo-ai-models" }, { "date": "2021-10-20", "date_label": "", "lab": "sk-hynix", "type": "chip", "title": "SK hynix develops HBM3", "summary": "SK hynix announced it had developed HBM3 memory, in 16 GB and 24 GB versions.", "primary_source": "https://www.thefpsreview.com/2021/10/20/sk-hynix-announces-hbm3-the-fastest-dram-in-the-world/", "secondary_source": "https://www.kedglobal.com/semiconductors/newsView/ked202110210003" }, { "date": "2021-11-08", "date_label": "", "lab": "amd", "type": "chip", "title": "AMD Instinct MI200 series (MI250X)", "summary": "AMD announced the Instinct MI200 series, including the MI250X, for exascale systems.", "primary_source": "https://ir.amd.com/news-events/press-releases/detail/1032/amd-unveils-workload-tailored-innovations-and-products-at-the-accelerated-data-center-premiere", "secondary_source": "" }, { "date": "2021-11-18", "date_label": "", "lab": "openai", "type": "api", "title": "OpenAI API waitlist removed", "summary": "OpenAI made its API available to developers in supported countries without a waitlist.", "primary_source": "https://openai.com/index/api-no-waitlist/", "secondary_source": "https://en.wikipedia.org/wiki/OpenAI" }, { "date": "2022-01-24", "date_label": "", "lab": "meta", "type": "infrastructure", "title": "Meta AI Research SuperCluster", "summary": "Meta introduced its AI Research SuperCluster, an AI supercomputer.", "primary_source": "https://about.fb.com/news/2022/01/introducing-metas-next-gen-ai-supercomputer/", "secondary_source": "" }, { "date": "2022-01-27", "date_label": "", "lab": "openai", "type": "model", "title": "InstructGPT becomes API default", "summary": "OpenAI made instruction-following models trained with human feedback (RLHF) the default in its API.", "primary_source": "https://openai.com/index/instruction-following/", "secondary_source": "https://en.wikipedia.org/wiki/GPT-3" }, { "date": "2022-01-27", "date_label": "", "lab": "jedec", "type": "chip", "title": "JEDEC publishes HBM3 standard", "summary": "JEDEC published the HBM3 standard, defining data rates up to 6.4 gigabits per second per pin, or 819 GB per second per device.", "primary_source": "https://www.jedec.org/news/pressreleases/jedec-publishes-hbm3-update-high-bandwidth-memory-hbm-standard", "secondary_source": "" }, { "date": "2022-02-07", "date_label": "", "lab": "nvidia", "type": "finance", "title": "Nvidia and SoftBank drop the Arm deal", "summary": "Nvidia and SoftBank ended the Arm purchase, citing regulatory obstacles, and SoftBank said it would explore a public offering of Arm.", "primary_source": "https://nvidianews.nvidia.com/news/nvidia-and-softbank-group-announce-termination-of-nvidias-acquisition-of-arm-limited", "secondary_source": "https://www.axios.com/2022/02/08/nvidia-arm-acquisition-chips" }, { "date": "2022-02-14", "date_label": "", "lab": "amd", "type": "chip", "title": "AMD completes Xilinx purchase", "summary": "AMD completed its all-stock acquisition of Xilinx, adding adaptive chips to its data centre portfolio.", "primary_source": "https://ir.amd.com/news-events/press-releases/detail/1047/amd-completes-acquisition-of-xilinx", "secondary_source": "" }, { "date": "2022-03-08", "date_label": "", "lab": "apple", "type": "chip", "title": "Apple M1 Ultra", "summary": "Apple unveiled the M1 Ultra, with up to 128 GB of unified memory.", "primary_source": "https://www.apple.com/newsroom/2022/03/apple-unveils-m1-ultra-the-worlds-most-powerful-chip-for-a-personal-computer/", "secondary_source": "" }, { "date": "2022-03-22", "date_label": "", "lab": "nvidia", "type": "chip", "title": "Nvidia Hopper and the H100", "summary": "Nvidia announced the Hopper architecture and the H100 GPU, with 80 billion transistors on TSMC's 4N process.", "primary_source": "https://nvidianews.nvidia.com/news/nvidia-announces-hopper-architecture-the-next-generation-of-accelerated-computing", "secondary_source": "" }, { "date": "2022-04-04", "date_label": "", "lab": "google", "type": "model", "title": "PaLM announced", "summary": "Google described PaLM, a 540-billion-parameter language model trained with its Pathways system.", "primary_source": "https://arxiv.org/abs/2204.02311", "secondary_source": "https://en.wikipedia.org/wiki/PaLM" }, { "date": "2022-04-06", "date_label": "", "lab": "openai", "type": "model", "title": "DALL-E 2 announced", "summary": "OpenAI introduced DALL-E 2, producing more realistic, higher-resolution images and image edits from text.", "primary_source": "https://openai.com/index/dall-e-2/", "secondary_source": "https://en.wikipedia.org/wiki/DALL-E" }, { "date": "2022-05-10", "date_label": "", "lab": "intel", "type": "chip", "title": "Intel's Habana launches Gaudi2", "summary": "Habana Labs launched the 7 nm Gaudi2 training chip at Intel Vision, with 96 GB of on-board memory and 24 Ethernet ports.", "primary_source": "https://habana.ai/press/habana-labs-launches-second-generation-ai-deep-learning-processors/", "secondary_source": "https://www.intc.com/news-events/press-releases/detail/1543/intel-announces-new-cloud-to-edge-technologies-to-solve" }, { "date": "2022-05-23", "date_label": "", "lab": "google", "type": "paper", "title": "Imagen text-to-image research", "summary": "Google Research presented Imagen, a text-to-image diffusion model, without public release.", "primary_source": "https://arxiv.org/abs/2205.11487", "secondary_source": "https://en.wikipedia.org/wiki/Imagen_(text-to-image_model)" }, { "date": "2022-05-30", "date_label": "", "lab": "us-doe", "type": "infrastructure", "title": "Frontier breaks the exascale barrier", "summary": "Oak Ridge's Frontier, built on AMD EPYC processors and Instinct accelerators, became the first system on the TOP500 to pass one exaflop, at 1.1 exaflops.", "primary_source": "https://www.ornl.gov/news/frontier-supercomputer-debuts-worlds-fastest-breaking-exascale-barrier", "secondary_source": "https://top500.org/lists/top500/2022/06/" }, { "date": "2022-06-13", "date_label": "", "lab": "eurohpc", "type": "hardware", "title": "LUMI inaugurated in Finland", "summary": "EuroHPC inaugurated LUMI in Kajaani, ranked third on the TOP500.", "primary_source": "https://www.eurohpc-ju.europa.eu/inauguration-lumi-fastest-greenest-supercomputer-europe-2022-06-13_en", "secondary_source": "https://www.hpcwire.com/2022/06/13/as-lumi-launches-a-look-at-the-state-of-eurohpc/" }, { "date": "2022-07-12", "date_label": "", "lab": "midjourney", "type": "product", "title": "Midjourney open beta", "summary": "Midjourney opened its Discord-based image generator to the public.", "primary_source": "", "secondary_source": "https://en.wikipedia.org/wiki/Midjourney" }, { "date": "2022-07-20", "date_label": "", "lab": "openai", "type": "product", "title": "DALL-E 2 enters beta", "summary": "OpenAI began inviting a million waitlisted users to DALL-E 2 with a credit-based pricing model.", "primary_source": "https://openai.com/index/dall-e-now-available-in-beta/", "secondary_source": "https://en.wikipedia.org/wiki/DALL-E" }, { "date": "2022-08-02", "date_label": "", "lab": "cxl", "type": "infrastructure", "title": "CXL 3.0 specification released", "summary": "The CXL Consortium released version 3.0, adding fabric capabilities and better memory sharing and pooling.", "primary_source": "https://www.hpcwire.com/off-the-wire/cxl-consortium-releases-compute-express-link-3-0-specification/", "secondary_source": "https://www.businesswire.com/news/home/20220802005028/en" }, { "date": "2022-08-09", "date_label": "", "lab": "us-gov", "type": "policy", "title": "CHIPS and Science Act signed", "summary": "The United States enacted the CHIPS and Science Act, funding domestic semiconductor manufacturing and research.", "primary_source": "https://bidenwhitehouse.archives.gov/briefing-room/statements-releases/2022/08/09/fact-sheet-chips-and-science-act-will-lower-costs-create-jobs-strengthen-supply-chains-and-counter-china/", "secondary_source": "" }, { "date": "2022-08-16", "date_label": "", "lab": "broadcom", "type": "chip", "title": "Broadcom ships Tomahawk 5 at 51.2 Tb/s", "summary": "Broadcom delivered a single Ethernet switch chip with 51.2 Tb/s of capacity, double any other on the market, by its account.", "primary_source": "https://investors.broadcom.com/news-releases/news-release-details/broadcom-ships-tomahawk-5-industrys-highest-bandwidth-switch", "secondary_source": "" }, { "date": "2022-08-22", "date_label": "", "lab": "stability", "type": "model", "title": "Stable Diffusion public release", "summary": "Stability AI and partners released Stable Diffusion with open weights.", "primary_source": "https://stability.ai/news/stable-diffusion-public-release", "secondary_source": "https://en.wikipedia.org/wiki/Stable_Diffusion" }, { "date": "2022-08-26", "date_label": "", "lab": "nvidia", "type": "policy", "title": "US imposes licence on Nvidia A100 and H100", "summary": "Nvidia disclosed that the US government had imposed a licence requirement on exports of A100 and H100 chips to China and Russia.", "primary_source": "https://www.sec.gov/Archives/edgar/data/1045810/000104581022000146/nvda-20220826.htm", "secondary_source": "" }, { "date": "2022-09-16", "date_label": "", "lab": "character-ai", "type": "product", "title": "Character.AI public beta", "summary": "Character.AI opened a public beta for chatting with user-created characters.", "primary_source": "https://beta.character.ai/", "secondary_source": "https://en.wikipedia.org/wiki/Character.ai" }, { "date": "2022-09-20", "date_label": "", "lab": "nvidia", "type": "chip", "title": "Nvidia H100 enters full production", "summary": "Nvidia said the H100 was in full production, with partners planning first products and services in October.", "primary_source": "https://nvidianews.nvidia.com/news/nvidia-hopper-in-full-production", "secondary_source": "" }, { "date": "2022-09-20", "date_label": "", "lab": "nvidia", "type": "chip", "title": "GeForce RTX 40 series (Ada Lovelace)", "summary": "Nvidia announced the GeForce RTX 40 series, led by the RTX 4090, on its Ada Lovelace architecture.", "primary_source": "https://nvidianews.nvidia.com/news/nvidia-introduces-geforce-rtx-40-series-gpus", "secondary_source": "" }, { "date": "2022-09-21", "date_label": "", "lab": "openai", "type": "model", "title": "Whisper speech recognition open-sourced", "summary": "OpenAI released Whisper, an open-source multilingual speech recognition model.", "primary_source": "https://openai.com/index/whisper/", "secondary_source": "https://en.wikipedia.org/wiki/Whisper_(speech_recognition_system)" }, { "date": "2022-09-28", "date_label": "", "lab": "openai", "type": "product", "title": "DALL-E 2 waitlist removed", "summary": "OpenAI opened DALL-E 2 to everyone without a waitlist.", "primary_source": "https://openai.com/index/dall-e-now-available-without-waitlist/", "secondary_source": "https://en.wikipedia.org/wiki/DALL-E" }, { "date": "2022-10-07", "date_label": "", "lab": "us-commerce", "type": "policy", "title": "US export controls on advanced chips, October 2022", "summary": "The US Commerce Department issued rules restricting exports of advanced computing chips and chipmaking equipment to China.", "primary_source": "https://www.federalregister.gov/documents/2022/10/13/2022-21658/implementation-of-additional-export-controls-certain-advanced-computing-and-semiconductor", "secondary_source": "" }, { "date": "2022-10-10", "date_label": "", "lab": "amazon", "type": "chip", "title": "AWS Trn1 instances generally available", "summary": "AWS made EC2 Trn1 instances, built on Trainium, generally available.", "primary_source": "https://aws.amazon.com/blogs/aws/amazon-ec2-trn1-instances-for-high-performance-model-training-are-now-available/", "secondary_source": "" }, { "date": "2022-10-13", "date_label": "", "lab": "us-bis", "type": "policy", "title": "US publishes first export limits on advanced AI chips", "summary": "BIS published rules requiring licences for advanced computing chips above set speed and data-rate limits bound for China.", "primary_source": "https://www.federalregister.gov/documents/2022/10/13/2022-21658/implementation-of-additional-export-controls-certain-advanced-computing-and-semiconductor", "secondary_source": "https://www.sec.gov/Archives/edgar/data/0001045810/000104581026000075/nvda-20260726.htm" }, { "date": "2022-11-05", "date_label": "", "lab": "midjourney", "type": "model", "title": "Midjourney V4 alpha", "summary": "Midjourney released its V4 model in alpha.", "primary_source": "", "secondary_source": "https://en.wikipedia.org/wiki/Midjourney" }, { "date": "2022-11-30", "date_label": "", "lab": "openai", "type": "product", "title": "ChatGPT launched as a research preview", "summary": "OpenAI released ChatGPT, a free conversational interface built on a GPT-3.5 model fine-tuned for dialogue.", "primary_source": "https://openai.com/index/chatgpt/", "secondary_source": "https://en.wikipedia.org/wiki/ChatGPT" }, { "date": "2022-12-15", "date_label": "", "lab": "anthropic", "type": "paper", "title": "Constitutional AI paper", "summary": "Anthropic described training a harmless assistant using a written set of principles and AI feedback.", "primary_source": "https://arxiv.org/abs/2212.08073", "secondary_source": "https://en.wikipedia.org/wiki/Claude_(language_model)" }, { "date": "2022-12-29", "date_label": "", "lab": "tsmc", "type": "chip", "title": "TSMC starts 3 nm volume production", "summary": "TSMC held a ceremony marking 3-nanometre volume production at Fab 18 in southern Taiwan.", "primary_source": "https://pr.tsmc.com/english/news/2986", "secondary_source": "" }, { "date": "2023-01-04", "date_label": "", "lab": "amd", "type": "chip", "title": "AMD brings Ryzen AI to laptop chips", "summary": "At CES 2023, AMD announced Ryzen 7000 mobile processors that bring Ryzen AI technology to select laptops.", "primary_source": "https://www.amd.com/en/newsroom/press-releases/2023-1-4-amd-extends-its-leadership-with-the-introduction-o.html", "secondary_source": "https://technalysisresearch.com/blogs/2023%20Blogs/december_15_2023_blog.html" }, { "date": "2023-01-23", "date_label": "", "lab": "openai", "type": "event", "title": "Microsoft extends OpenAI partnership", "summary": "Microsoft announced a multiyear, multibillion-dollar extension of its investment in OpenAI.", "primary_source": "https://openai.com/index/openai-and-microsoft-extend-partnership/", "secondary_source": "https://en.wikipedia.org/wiki/OpenAI" }, { "date": "2023-02-01", "date_label": "", "lab": "openai", "type": "product", "title": "ChatGPT Plus subscription", "summary": "OpenAI introduced a $20 monthly plan with priority access and faster responses.", "primary_source": "https://openai.com/index/chatgpt-plus/", "secondary_source": "https://en.wikipedia.org/wiki/ChatGPT" }, { "date": "2023-02-03", "date_label": "", "lab": "quora", "type": "product", "title": "Poe opens to the public", "summary": "Quora opened Poe, an app for chatting with several AI bots, to everyone on iOS.", "primary_source": "https://apps.apple.com/us/app/poe-fast-helpful-ai-chat/id1640745955", "secondary_source": "https://techcrunch.com/2023/02/06/quora-opens-its-new-ai-chatbot-app-poe-to-the-general-public/" }, { "date": "2023-02-06", "date_label": "", "lab": "google", "type": "product", "title": "Bard announced", "summary": "Google announced Bard, an experimental conversational AI service based on LaMDA, for trusted testers.", "primary_source": "https://blog.google/technology/ai/bard-google-ai-search-updates/", "secondary_source": "https://en.wikipedia.org/wiki/Gemini_(chatbot)" }, { "date": "2023-02-07", "date_label": "", "lab": "microsoft", "type": "product", "title": "New Bing with chat", "summary": "Microsoft unveiled an AI-powered Bing and Edge with a chat assistant built on an OpenAI model.", "primary_source": "https://blogs.microsoft.com/blog/2023/02/07/reinventing-search-with-a-new-ai-powered-microsoft-bing-and-edge-your-copilot-for-the-web/", "secondary_source": "https://en.wikipedia.org/wiki/Microsoft_Copilot" }, { "date": "2023-02-24", "date_label": "", "lab": "meta", "type": "model", "title": "LLaMA released to researchers", "summary": "Meta released LLaMA, a family of 7B to 65B models, under a research-only licence; the weights soon leaked.", "primary_source": "https://ai.meta.com/blog/large-language-model-llama-meta-ai/", "secondary_source": "https://en.wikipedia.org/wiki/Llama_(language_model)" }, { "date": "2023-03-01", "date_label": "", "lab": "openai", "type": "api", "title": "ChatGPT (gpt-3.5-turbo) and Whisper APIs", "summary": "OpenAI opened the gpt-3.5-turbo chat model and Whisper to developers at much lower prices than earlier models.", "primary_source": "https://openai.com/index/introducing-chatgpt-and-whisper-apis/", "secondary_source": "https://en.wikipedia.org/wiki/ChatGPT" }, { "date": "2023-03-10", "date_label": "", "lab": "community", "type": "product", "title": "llama.cpp released", "summary": "Georgi Gerganov released llama.cpp, letting LLaMA models run on ordinary laptops in C/C++.", "primary_source": "https://github.com/ggml-org/llama.cpp", "secondary_source": "https://en.wikipedia.org/wiki/Llama.cpp" }, { "date": "2023-03-14", "date_label": "", "lab": "openai", "type": "model", "title": "GPT-4 released", "summary": "OpenAI announced GPT-4, a large multimodal model, available in ChatGPT Plus and via an API waitlist.", "primary_source": "https://openai.com/index/gpt-4-research/", "secondary_source": "https://en.wikipedia.org/wiki/GPT-4" }, { "date": "2023-03-14", "date_label": "", "lab": "anthropic", "type": "model", "title": "Claude launched", "summary": "Anthropic released Claude and Claude Instant to businesses through its API and partners.", "primary_source": "https://www.anthropic.com/news/introducing-claude", "secondary_source": "https://www.theverge.com/2023/3/14/23640056/anthropic-ai-chatbot-claude-google-launch" }, { "date": "2023-03-15", "date_label": "", "lab": "openai", "type": "paper", "title": "GPT-4 Technical Report on arXiv", "summary": "OpenAI posted the GPT-4 technical report, giving benchmark results but few architecture details.", "primary_source": "https://arxiv.org/abs/2303.08774", "secondary_source": "https://en.wikipedia.org/wiki/GPT-4" }, { "date": "2023-03-15", "date_label": "", "lab": "midjourney", "type": "model", "title": "Midjourney V5", "summary": "Midjourney released V5 in alpha with more photorealistic output.", "primary_source": "", "secondary_source": "https://en.wikipedia.org/wiki/Midjourney" }, { "date": "2023-03-16", "date_label": "", "lab": "microsoft", "type": "product", "title": "Microsoft 365 Copilot announced", "summary": "Microsoft announced Copilot for Word, Excel, PowerPoint, Outlook and Teams.", "primary_source": "https://blogs.microsoft.com/blog/2023/03/16/introducing-microsoft-365-copilot-your-copilot-for-work/", "secondary_source": "https://en.wikipedia.org/wiki/Microsoft_Copilot" }, { "date": "2023-03-21", "date_label": "", "lab": "google", "type": "product", "title": "Bard opens to US and UK users", "summary": "Google began giving users in the US and UK access to Bard through a waitlist.", "primary_source": "https://blog.google/technology/ai/try-bard/", "secondary_source": "https://en.wikipedia.org/wiki/Gemini_(chatbot)" }, { "date": "2023-03-23", "date_label": "", "lab": "openai", "type": "feature", "title": "ChatGPT plugins announced", "summary": "OpenAI introduced plugins, including web browsing and a code interpreter, to a limited set of users.", "primary_source": "https://openai.com/index/chatgpt-plugins/", "secondary_source": "https://en.wikipedia.org/wiki/ChatGPT" }, { "date": "2023-03-30", "date_label": "", "lab": "community", "type": "product", "title": "AutoGPT released", "summary": "Toran Bruce Richards released AutoGPT, an open-source agent that chains GPT-4 calls toward goals.", "primary_source": "https://github.com/Significant-Gravitas/AutoGPT", "secondary_source": "https://en.wikipedia.org/wiki/AutoGPT" }, { "date": "2023-04-05", "date_label": "", "lab": "google", "type": "infrastructure", "title": "Google details TPU v4 optical switching", "summary": "Google published details on TPU v4 pods, joined by its own optical circuit switches, and said the chip beat TPU v3 by 2.1 times per chip.", "primary_source": "https://cloud.google.com/blog/topics/systems/tpu-v4-enables-performance-energy-and-co2e-efficiency-gains", "secondary_source": "" }, { "date": "2023-04-11", "date_label": "", "lab": "alibaba", "type": "product", "title": "Alibaba unveils Tongyi Qianwen", "summary": "Alibaba Cloud announced its Tongyi Qianwen (Qwen) chatbot model for business testing.", "primary_source": "https://www.aap.com.au/aapreleases/cision20230410ae66473", "secondary_source": "https://www.reuters.com/technology/alibaba-unveils-tongyi-qianwen-an-ai-model-similar-gpt-2023-04-11/" }, { "date": "2023-04-13", "date_label": "", "lab": "amazon", "type": "chip", "title": "AWS Inf2 instances generally available", "summary": "AWS made EC2 Inf2 instances, built on Inferentia2, generally available.", "primary_source": "https://aws.amazon.com/blogs/aws/amazon-ec2-inf2-instances-for-low-cost-high-performance-generative-ai-inference-are-now-generally-available/", "secondary_source": "" }, { "date": "2023-04-25", "date_label": "", "lab": "openai", "type": "feature", "title": "Option to turn off chat history", "summary": "ChatGPT users gained a setting to stop conversations being used for training and kept in history.", "primary_source": "https://openai.com/index/new-ways-to-manage-your-data-in-chatgpt/", "secondary_source": "https://en.wikipedia.org/wiki/ChatGPT" }, { "date": "2023-05-10", "date_label": "", "lab": "google", "type": "model", "title": "PaLM 2 at Google I/O", "summary": "Google announced PaLM 2, powering Bard and many Google products, and opened Bard without a waitlist.", "primary_source": "https://blog.google/technology/ai/google-palm-2-ai-large-language-model/", "secondary_source": "https://en.wikipedia.org/wiki/PaLM" }, { "date": "2023-05-11", "date_label": "", "lab": "anthropic", "type": "feature", "title": "100K context window", "summary": "Anthropic expanded Claude's context window to 100,000 tokens.", "primary_source": "https://www.anthropic.com/news/100k-context-windows", "secondary_source": "https://en.wikipedia.org/wiki/Claude_(language_model)" }, { "date": "2023-05-12", "date_label": "", "lab": "openai", "type": "feature", "title": "Web browsing and plugins beta for Plus", "summary": "OpenAI began rolling out web browsing and plugins to all ChatGPT Plus subscribers as beta features.", "primary_source": "https://help.openai.com/en/articles/6825453-chatgpt-release-notes", "secondary_source": "https://en.wikipedia.org/wiki/ChatGPT" }, { "date": "2023-05-18", "date_label": "", "lab": "openai", "type": "product", "title": "ChatGPT iOS app", "summary": "OpenAI released an official ChatGPT app for iPhone in the US.", "primary_source": "https://openai.com/index/introducing-the-chatgpt-app-for-ios/", "secondary_source": "https://en.wikipedia.org/wiki/ChatGPT" }, { "date": "2023-05-18", "date_label": "", "lab": "meta", "type": "chip", "title": "Meta MTIA v1", "summary": "Meta described MTIA v1, an in-house inference accelerator made on TSMC's 7-nanometre process.", "primary_source": "https://ai.meta.com/blog/meta-training-inference-accelerator-AI-MTIA/", "secondary_source": "" }, { "date": "2023-05-28", "date_label": "", "lab": "nvidia", "type": "hardware", "title": "Grace Hopper in production and the DGX GH200", "summary": "Nvidia announced the DGX GH200 AI supercomputer, powered by its GH200 Grace Hopper Superchips.", "primary_source": "https://nvidianews.nvidia.com/news/nvidia-announces-dgx-gh200-ai-supercomputer", "secondary_source": "https://nvidianews.nvidia.com/news/nvidia-grace-hopper-superchips-designed-for-accelerated-generative-ai-enter-full-production" }, { "date": "2023-05-28", "date_label": "", "lab": "nvidia", "type": "infrastructure", "title": "Nvidia Spectrum-X Ethernet for AI", "summary": "Nvidia launched Spectrum-X, an Ethernet networking platform designed for generative AI data centres.", "primary_source": "https://nvidianews.nvidia.com/news/nvidia-launches-accelerated-ethernet-platform-for-hyperscale-generative-ai", "secondary_source": "" }, { "date": "2023-06-05", "date_label": "", "lab": "apple", "type": "chip", "title": "Apple M2 Ultra", "summary": "Apple introduced the M2 Ultra, with up to 192 GB of unified memory and 800 GB per second of bandwidth.", "primary_source": "https://www.apple.com/newsroom/2023/06/apple-introduces-m2-ultra/", "secondary_source": "" }, { "date": "2023-06-13", "date_label": "", "lab": "openai", "type": "api", "title": "Function calling in the API", "summary": "OpenAI added function calling and new GPT-4 and GPT-3.5-turbo versions with lower prices.", "primary_source": "https://openai.com/index/function-calling-and-other-api-updates/", "secondary_source": "https://en.wikipedia.org/wiki/OpenAI" }, { "date": "2023-06-30", "date_label": "", "lab": "netherlands", "type": "policy", "title": "Dutch controls on chipmaking equipment", "summary": "The Dutch government published export controls requiring a licence for advanced chipmaking equipment, effective September 1, 2023.", "primary_source": "https://www.aljazeera.com/news/2023/6/30/dutch-firms-must-seek-licence-before-exporting-chip-gear-govt", "secondary_source": "https://www.hlc.com/en/publications/new-dutch-export-controls-on-advanced-semiconductor-manufacturing-equipment" }, { "date": "2023-07-03", "date_label": "", "lab": "china", "type": "policy", "title": "China announces gallium and germanium export controls", "summary": "China's commerce ministry and customs announced export licences for gallium, germanium and related compounds, from August 1.", "primary_source": "https://www.globalpolicywatch.com/?p=13992", "secondary_source": "https://www.globalsecurity.org/wmd/library/news/china/2023/07/china-230703-globaltimes06.htm" }, { "date": "2023-07-06", "date_label": "", "lab": "openai", "type": "api", "title": "GPT-4 API generally available; Code Interpreter rolls out", "summary": "OpenAI opened GPT-4 to all paying API users and began rolling out Code Interpreter to ChatGPT Plus.", "primary_source": "https://openai.com/index/gpt-4-api-general-availability/", "secondary_source": "https://help.openai.com/en/articles/6825453-chatgpt-release-notes" }, { "date": "2023-07-08", "date_label": "", "lab": "community", "type": "product", "title": "Ollama first release", "summary": "Ollama released a tool for downloading and running open models locally.", "primary_source": "https://github.com/ollama/ollama", "secondary_source": "https://en.wikipedia.org/wiki/Ollama" }, { "date": "2023-07-11", "date_label": "", "lab": "anthropic", "type": "model", "title": "Claude 2 and claude.ai", "summary": "Anthropic released Claude 2 and opened the claude.ai chat site to the public in the US and UK.", "primary_source": "https://www.anthropic.com/news/claude-2", "secondary_source": "https://en.wikipedia.org/wiki/Claude_(language_model)" }, { "date": "2023-07-18", "date_label": "", "lab": "meta", "type": "model", "title": "Llama 2", "summary": "Meta released Llama 2 with a licence allowing commercial use, in partnership with Microsoft.", "primary_source": "https://about.fb.com/news/2023/07/llama-2/", "secondary_source": "https://en.wikipedia.org/wiki/Llama_(language_model)" }, { "date": "2023-07-19", "date_label": "", "lab": "uec", "type": "infrastructure", "title": "Ultra Ethernet Consortium founded", "summary": "AMD, Arista, Broadcom, Cisco, Eviden, HPE, Intel, Meta and Microsoft formed a Linux Foundation group to adapt Ethernet to AI and high-performance computing.", "primary_source": "https://www.linuxfoundation.org/press/announcing-ultra-ethernet-consortium-uec", "secondary_source": "https://www.sdxcentral.com/articles/news/ultra-ethernet-consortium-focuses-on-ai-and-hpc-takes-aim-at-infiniband/2023/07/" }, { "date": "2023-07-20", "date_label": "", "lab": "openai", "type": "feature", "title": "Custom instructions in ChatGPT", "summary": "ChatGPT gained custom instructions so users can set preferences that apply to every conversation.", "primary_source": "https://openai.com/index/custom-instructions-for-chatgpt/", "secondary_source": "https://en.wikipedia.org/wiki/ChatGPT" }, { "date": "2023-08-03", "date_label": "", "lab": "alibaba", "type": "model", "title": "Qwen-7B open weights", "summary": "Alibaba released Qwen-7B as open weights.", "primary_source": "https://github.com/QwenLM/Qwen-7B", "secondary_source": "https://www.reuters.com/technology/alibaba-unveils-open-sourced-ai-model-similar-metas-llama-2-2023-08-03/" }, { "date": "2023-08-21", "date_label": "", "lab": "sk-hynix", "type": "chip", "title": "SK hynix HBM3E samples", "summary": "SK hynix said it had developed HBM3E and was providing samples to a customer for evaluation.", "primary_source": "https://news.skhynix.com/en/sk-hynix-develops-worlds-best-performing-hbm3e/", "secondary_source": "" }, { "date": "2023-08-22", "date_label": "", "lab": "openai", "type": "api", "title": "Fine-tuning for GPT-3.5 Turbo", "summary": "Developers could fine-tune GPT-3.5 Turbo on their own data.", "primary_source": "https://openai.com/index/gpt-3-5-turbo-fine-tuning-and-api-updates/", "secondary_source": "https://en.wikipedia.org/wiki/OpenAI" }, { "date": "2023-08-24", "date_label": "", "lab": "meta", "type": "model", "title": "Code Llama", "summary": "Meta released Code Llama, Llama 2 models specialized for programming.", "primary_source": "https://ai.meta.com/blog/code-llama-large-language-model-coding/", "secondary_source": "https://en.wikipedia.org/wiki/Llama_(language_model)" }, { "date": "2023-08-28", "date_label": "", "lab": "openai", "type": "product", "title": "ChatGPT Enterprise", "summary": "OpenAI launched a business plan with enterprise security, unlimited GPT-4 and admin tools.", "primary_source": "https://openai.com/index/introducing-chatgpt-enterprise/", "secondary_source": "https://en.wikipedia.org/wiki/ChatGPT" }, { "date": "2023-08-29", "date_label": "", "lab": "google", "type": "chip", "title": "Google Cloud TPU v5e", "summary": "Google announced Cloud TPU v5e, a chip aimed at cost-efficient training and inference.", "primary_source": "https://cloud.google.com/blog/products/compute/announcing-cloud-tpu-v5e-and-a3-gpus-in-ga", "secondary_source": "" }, { "date": "2023-09-01", "date_label": "", "lab": "netherlands", "type": "policy", "title": "Dutch chip-equipment licence rule takes effect", "summary": "The Dutch national authorisation requirement for certain advanced chipmaking equipment entered into force.", "primary_source": "https://web.archive.org/web/20240615150738id_/https://www.government.nl/latest/news/2023/06/30/government-publishes-additional-export-measures-for-advanced-semiconductor-manufacturing-equipment", "secondary_source": "https://hlc.com/en/publications/new-dutch-export-controls-on-advanced-semiconductor-manufacturing-equipment" }, { "date": "2023-09-20", "date_label": "", "lab": "openai", "type": "model", "title": "DALL-E 3 announced", "summary": "OpenAI announced DALL-E 3, built to follow prompts closely and to be used through ChatGPT.", "primary_source": "https://openai.com/index/dall-e-3/", "secondary_source": "https://en.wikipedia.org/wiki/DALL-E" }, { "date": "2023-09-21", "date_label": "", "lab": "eu", "type": "policy", "title": "European Chips Act enters into force", "summary": "The European Chips Act took effect, with a goal of doubling the EU's global chip market share to 20% by 2030.", "primary_source": "https://digital-strategy.ec.europa.eu/en/news/digital-sovereignty-european-chips-act-enters-force", "secondary_source": "https://ec.europa.eu/commission/presscorner/api/files/document/print/en/ip_23_4518/IP_23_4518_EN.pdf" }, { "date": "2023-09-25", "date_label": "", "lab": "openai", "type": "feature", "title": "ChatGPT can see, hear and speak", "summary": "OpenAI announced voice conversations and image input (GPT-4V) for ChatGPT Plus and Enterprise.", "primary_source": "https://openai.com/index/chatgpt-can-now-see-hear-and-speak/", "secondary_source": "https://en.wikipedia.org/wiki/GPT-4" }, { "date": "2023-09-25", "date_label": "", "lab": "anthropic", "type": "finance", "title": "Amazon to invest up to $4 billion in Anthropic", "summary": "Amazon agreed to invest up to $4 billion in Anthropic for a minority stake.", "primary_source": "https://www.aboutamazon.com/news/company-news/amazon-aws-anthropic-ai", "secondary_source": "https://www.sec.gov/Archives/edgar/data/2130357/000162828026067168/anthropicpbc-sx1.htm" }, { "date": "2023-09-27", "date_label": "", "lab": "mistral", "type": "model", "title": "Mistral 7B", "summary": "Mistral AI released its first model, Mistral 7B, under Apache 2.0.", "primary_source": "https://mistral.ai/news/announcing-mistral-7b/", "secondary_source": "https://en.wikipedia.org/wiki/Mistral_AI" }, { "date": "2023-10-25", "date_label": "", "lab": "us-commerce", "type": "policy", "title": "US tightens chip export controls, October 2023", "summary": "The Federal Register published US Commerce rules tightening controls on advanced computing chips and chipmaking equipment.", "primary_source": "https://www.federalregister.gov/documents/2023/10/25/2023-23055/implementation-of-additional-export-controls-certain-advanced-computing-items-supercomputer-and", "secondary_source": "" }, { "date": "2023-10-30", "date_label": "", "lab": "apple", "type": "chip", "title": "Apple M3 family", "summary": "Apple unveiled the M3, M3 Pro and M3 Max, with up to 128 GB of unified memory.", "primary_source": "https://www.apple.com/newsroom/2023/10/apple-unveils-m3-m3-pro-and-m3-max-the-most-advanced-chips-for-a-personal-computer/", "secondary_source": "" }, { "date": "2023-11-04", "date_label": "", "lab": "xai", "type": "model", "title": "Grok announced", "summary": "xAI unveiled Grok, a chatbot for X Premium+ subscribers.", "primary_source": "https://x.ai/news/grok", "secondary_source": "https://en.wikipedia.org/wiki/Grok_(chatbot)" }, { "date": "2023-11-06", "date_label": "", "lab": "openai", "type": "api", "title": "Assistants API beta", "summary": "OpenAI released a beta API for building assistants with threads, retrieval and code interpreter tools.", "primary_source": "https://openai.com/index/new-models-and-developer-products-announced-at-devday/", "secondary_source": "https://en.wikipedia.org/wiki/OpenAI" }, { "date": "2023-11-06", "date_label": "", "lab": "openai", "type": "event", "title": "OpenAI DevDay 2023", "summary": "At its first developer conference OpenAI announced GPT-4 Turbo with 128K context, the Assistants API, GPTs and lower prices.", "primary_source": "https://openai.com/index/new-models-and-developer-products-announced-at-devday/", "secondary_source": "https://en.wikipedia.org/wiki/OpenAI" }, { "date": "2023-11-06", "date_label": "", "lab": "openai", "type": "model", "title": "GPT-4 Turbo preview", "summary": "OpenAI previewed GPT-4 Turbo with a 128,000-token context window and cheaper pricing.", "primary_source": "https://openai.com/index/new-models-and-developer-products-announced-at-devday/", "secondary_source": "https://en.wikipedia.org/wiki/GPT-4" }, { "date": "2023-11-06", "date_label": "", "lab": "openai", "type": "feature", "title": "GPTs: custom versions of ChatGPT", "summary": "OpenAI let users build custom ChatGPT versions with instructions, knowledge files and actions, without code.", "primary_source": "https://openai.com/index/introducing-gpts/", "secondary_source": "https://en.wikipedia.org/wiki/GPT_Store" }, { "date": "2023-11-13", "date_label": "", "lab": "nvidia", "type": "chip", "title": "Nvidia H200", "summary": "Nvidia announced the H200, which it described as the first GPU with HBM3e memory.", "primary_source": "https://nvidianews.nvidia.com/news/nvidia-supercharges-hopper-the-worlds-leading-ai-computing-platform", "secondary_source": "" }, { "date": "2023-11-15", "date_label": "", "lab": "microsoft", "type": "chip", "title": "Microsoft Maia 100", "summary": "Microsoft introduced Maia 100, its first custom AI accelerator, at Ignite.", "primary_source": "https://azure.microsoft.com/en-us/blog/microsoft-azure-delivers-purpose-built-cloud-infrastructure-in-the-era-of-ai/", "secondary_source": "https://news.microsoft.com/source/features/ai/in-house-chips-silicon-to-service-to-meet-ai-demand/" }, { "date": "2023-11-17", "date_label": "", "lab": "openai", "type": "event", "title": "Board removes Sam Altman as CEO", "summary": "OpenAI's board announced Sam Altman's departure as CEO; he returned days later after staff and investor pressure.", "primary_source": "https://openai.com/index/openai-announces-leadership-transition/", "secondary_source": "https://en.wikipedia.org/wiki/Removal_of_Sam_Altman_from_OpenAI" }, { "date": "2023-11-21", "date_label": "", "lab": "anthropic", "type": "model", "title": "Claude 2.1", "summary": "Anthropic released Claude 2.1 with a 200,000-token context window and beta tool use.", "primary_source": "https://www.anthropic.com/news/claude-2-1", "secondary_source": "https://www.theverge.com/2023/11/21/23971070/anthropic-claude-2-1-openai-ai-chatbot-update-beta-tools" }, { "date": "2023-11-22", "date_label": "", "lab": "openai", "type": "event", "title": "Sam Altman agreed to return as CEO", "summary": "OpenAI reached an agreement for Altman to return as CEO with a new initial board.", "primary_source": "https://openai.com/index/sam-altman-returns-as-ceo-openai-has-a-new-initial-board/", "secondary_source": "https://en.wikipedia.org/wiki/Removal_of_Sam_Altman_from_OpenAI" }, { "date": "2023-11-28", "date_label": "", "lab": "amazon", "type": "chip", "title": "AWS announces Trainium2", "summary": "AWS announced Trainium2, its second training chip, at re:Invent.", "primary_source": "https://press.aboutamazon.com/2023/11/aws-unveils-next-generation-aws-designed-chips", "secondary_source": "" }, { "date": "2023-12-05", "date_label": "", "lab": "jedec", "type": "hardware", "title": "JEDEC publishes the CAMM2 module standard", "summary": "JEDEC published JESD318, a compression-attached module standard covering DDR5 and LPDDR5/5X.", "primary_source": "https://www.jedec.org/news/pressreleases/jedec-publishes-new-camm2-memory-module-standard", "secondary_source": "" }, { "date": "2023-12-06", "date_label": "", "lab": "google", "type": "model", "title": "Gemini 1.0 announced", "summary": "Google DeepMind introduced Gemini in Ultra, Pro and Nano sizes; Pro began powering Bard the same day.", "primary_source": "https://blog.google/technology/ai/google-gemini-ai/", "secondary_source": "https://en.wikipedia.org/wiki/Gemini_(language_model)" }, { "date": "2023-12-06", "date_label": "", "lab": "google", "type": "chip", "title": "Google announces Cloud TPU v5p and AI Hypercomputer", "summary": "Google announced TPU v5p, with pods of 8,960 chips, which it said train large language models 2.8 times faster than TPU v4.", "primary_source": "https://cloud.google.com/blog/products/ai-machine-learning/introducing-cloud-tpu-v5p-and-ai-hypercomputer", "secondary_source": "https://datacenterdynamics.com/en/news/google-debuts-cloud-tpu-v5p-chip-and-ai-hypercomputer" }, { "date": "2023-12-06", "date_label": "", "lab": "amd", "type": "chip", "title": "AMD Instinct MI300X and MI300A", "summary": "AMD launched the MI300 series, including the MI300X with 192 GB of HBM3 memory and the MI300A, which combines CPU and GPU.", "primary_source": "https://www.amd.com/en/newsroom/press-releases/2023-12-6-amd-delivers-leadership-portfolio-of-data-center-a.html", "secondary_source": "" }, { "date": "2023-12-11", "date_label": "", "lab": "mistral", "type": "model", "title": "Mixtral 8x7B", "summary": "Mistral released Mixtral 8x7B, an open sparse mixture-of-experts model, after posting the weights by torrent.", "primary_source": "https://mistral.ai/news/mixtral-of-experts/", "secondary_source": "https://en.wikipedia.org/wiki/Mistral_AI" }, { "date": "2023-12-14", "date_label": "", "lab": "intel", "type": "chip", "title": "Intel launches Core Ultra with an NPU", "summary": "Intel launched Core Ultra, its first consumer chips with a neural processing unit, at its AI Everywhere event.", "primary_source": "https://www.intel.com/content/www/us/en/newsroom/news/core-ultra-client-computing-news-1.html", "secondary_source": "https://www.techradar.com/computing/cpu/intel-announces-new-core-ultra-cpu-with-ai-processing-engine-to-launch-this-year" }, { "date": "2023-12-21", "date_label": "", "lab": "midjourney", "type": "model", "title": "Midjourney V6", "summary": "Midjourney released V6 in alpha with better prompt following and text.", "primary_source": "", "secondary_source": "https://en.wikipedia.org/wiki/Midjourney" }, { "date": "2024-01-10", "date_label": "", "lab": "openai", "type": "product", "title": "GPT Store and ChatGPT Team", "summary": "OpenAI opened the GPT Store for sharing custom GPTs and launched the ChatGPT Team plan for small groups.", "primary_source": "https://openai.com/index/introducing-the-gpt-store/", "secondary_source": "https://en.wikipedia.org/wiki/GPT_Store" }, { "date": "2024-02-08", "date_label": "", "lab": "google", "type": "product", "title": "Bard renamed Gemini; Gemini Advanced", "summary": "Google rebranded Bard as Gemini and launched Gemini Advanced with Ultra 1.0 in a paid plan.", "primary_source": "https://blog.google/products/gemini/bard-gemini-advanced-app/", "secondary_source": "https://en.wikipedia.org/wiki/Gemini_(chatbot)" }, { "date": "2024-02-13", "date_label": "", "lab": "openai", "type": "feature", "title": "ChatGPT memory testing begins", "summary": "OpenAI started testing memory, letting ChatGPT remember details across conversations, with user controls.", "primary_source": "https://openai.com/index/memory-and-new-controls-for-chatgpt/", "secondary_source": "https://en.wikipedia.org/wiki/ChatGPT" }, { "date": "2024-02-15", "date_label": "", "lab": "openai", "type": "model", "title": "Sora text-to-video preview", "summary": "OpenAI previewed Sora, a model that makes up to minute-long videos from text, to red teamers and selected artists.", "primary_source": "https://openai.com/index/sora/", "secondary_source": "https://en.wikipedia.org/wiki/Sora_(text-to-video_model)" }, { "date": "2024-02-15", "date_label": "", "lab": "google", "type": "model", "title": "Gemini 1.5 Pro", "summary": "Google announced Gemini 1.5 Pro with a context window of up to one million tokens.", "primary_source": "https://blog.google/technology/ai/google-gemini-next-generation-model-february-2024/", "secondary_source": "https://en.wikipedia.org/wiki/Gemini_(language_model)" }, { "date": "2024-02-15", "date_label": "", "lab": "nvidia", "type": "infrastructure", "title": "Nvidia reveals Eos", "summary": "Nvidia gave a look at Eos, its in-house supercomputer, ranked in the world's top 10.", "primary_source": "https://blogs.nvidia.com/blog/eos/", "secondary_source": "" }, { "date": "2024-02-20", "date_label": "", "lab": "dipc", "type": "hardware", "title": "DIPC's Hyperion supercomputer shown in Spain", "summary": "A Basque minister visited Hyperion at the Donostia physics centre, a 2.8-million-euro system with over 14,000 cores.", "primary_source": "https://dipc.ehu.eus/en/whats-new/news/hyperion-the-new-supercomputer-at-dipc", "secondary_source": "https://news.lenovo.com/?p=27959" }, { "date": "2024-02-21", "date_label": "", "lab": "google", "type": "model", "title": "Gemma open models", "summary": "Google released Gemma, lightweight open-weight models built from Gemini research.", "primary_source": "https://blog.google/technology/developers/gemma-open-models/", "secondary_source": "https://en.wikipedia.org/wiki/Gemma_(language_model)" }, { "date": "2024-02-26", "date_label": "", "lab": "mistral", "type": "model", "title": "Mistral Large and Le Chat", "summary": "Mistral launched Mistral Large and its Le Chat assistant, plus a Microsoft partnership.", "primary_source": "https://mistral.ai/news/mistral-large/", "secondary_source": "https://en.wikipedia.org/wiki/Mistral_AI" }, { "date": "2024-02-26", "date_label": "", "lab": "micron", "type": "chip", "title": "Micron starts HBM3E volume production", "summary": "Micron said it had begun volume production of HBM3E memory.", "primary_source": "https://investors.micron.com/news/press-release/2024/Micron-Commences-Volume-Production-of-Industry-Leading-HBM3E-Solution-to-Accelerate-the-Growth-of-AI-02-26-2024/default.aspx", "secondary_source": "" }, { "date": "2024-02-27", "date_label": "", "lab": "samsung", "type": "chip", "title": "Samsung 36 GB HBM3E 12H", "summary": "Samsung announced what it called the industry's first 36 GB, 12-layer HBM3E.", "primary_source": "https://news.samsung.com/global/samsung-develops-industry-first-36gb-hbm3e-12h-dram", "secondary_source": "" }, { "date": "2024-03-04", "date_label": "", "lab": "anthropic", "type": "model", "title": "Claude 3 family", "summary": "Anthropic released Claude 3 Opus, Sonnet and Haiku, adding image input.", "primary_source": "https://www.anthropic.com/news/claude-3-family", "secondary_source": "https://www.reuters.com/technology/anthropic-releases-more-powerful-claude-3-ai-tech-race-continues-2024-03-04/" }, { "date": "2024-03-05", "date_label": "", "lab": "jedec", "type": "chip", "title": "JEDEC publishes GDDR7", "summary": "JEDEC published JESD239, the GDDR7 standard, which doubles GDDR6 bandwidth to as much as 192 GB/s per device.", "primary_source": "https://www.jedec.org/news/pressreleases/jedec-publishes-gddr7-graphics-memory-standard", "secondary_source": "" }, { "date": "2024-03-12", "date_label": "", "lab": "meta", "type": "infrastructure", "title": "Meta details 24,576-GPU clusters", "summary": "Meta shared details of two 24,576-GPU clusters supporting Llama 3 development, one on RoCE Ethernet and one on InfiniBand.", "primary_source": "https://engineering.fb.com/2024/03/12/data-center-engineering/building-metas-genai-infrastructure/", "secondary_source": "" }, { "date": "2024-03-13", "date_label": "", "lab": "cerebras", "type": "chip", "title": "Cerebras WSE-3", "summary": "Cerebras announced the 5-nanometre WSE-3, with 4 trillion transistors and 900,000 cores.", "primary_source": "https://www.cerebras.ai/press-release/cerebras-announces-third-generation-wafer-scale-engine", "secondary_source": "" }, { "date": "2024-03-17", "date_label": "", "lab": "xai", "type": "model", "title": "Grok-1 weights released", "summary": "xAI released the 314B-parameter Grok-1 base model weights under Apache 2.0.", "primary_source": "https://github.com/xai-org/grok-1", "secondary_source": "https://en.wikipedia.org/wiki/Grok_(chatbot)" }, { "date": "2024-03-18", "date_label": "", "lab": "nvidia", "type": "infrastructure", "title": "Nvidia announces X800 800 Gb/s switches", "summary": "Nvidia announced Quantum-X800 InfiniBand and Spectrum-X800 Ethernet, which it called the first end-to-end 800 Gb/s networking platforms.", "primary_source": "https://www.nvidia.com/ja-jp/about-nvidia/press-releases/2024/networking-switches-gpu-computing-ai", "secondary_source": "https://businessnetwork.jp/article/19526" }, { "date": "2024-03-18", "date_label": "", "lab": "nvidia", "type": "chip", "title": "Nvidia Blackwell, B200 and GB200", "summary": "Nvidia announced the Blackwell platform, with 208 billion transistors per GPU on TSMC's 4NP process, along with the B200 GPU and the GB200 Grace Blackwell Superchip.", "primary_source": "https://nvidianews.nvidia.com/news/nvidia-blackwell-platform-arrives-to-power-a-new-era-of-computing", "secondary_source": "" }, { "date": "2024-04-01", "date_label": "", "lab": "openai", "type": "product", "title": "ChatGPT without an account", "summary": "OpenAI began letting people use ChatGPT without signing up.", "primary_source": "https://openai.com/index/start-using-chatgpt-instantly/", "secondary_source": "https://en.wikipedia.org/wiki/ChatGPT" }, { "date": "2024-04-09", "date_label": "", "lab": "google", "type": "product", "title": "TPU v5p reaches general availability", "summary": "Google said at Next '24 that TPU v5p was generally available.", "primary_source": "https://cloud.google.com/blog/products/compute/whats-new-with-google-clouds-ai-hypercomputer-architecture", "secondary_source": "https://atmarkit.itmedia.co.jp/ait/articles/2404/10/news085.html" }, { "date": "2024-04-09", "date_label": "", "lab": "intel", "type": "chip", "title": "Intel Gaudi 3", "summary": "Intel introduced Gaudi 3 with 128 GB of HBM memory and, by its own projections, about 50 percent faster training than the H100 on selected models.", "primary_source": "https://www.intel.com/content/www/us/en/newsroom/news/vision-2024-gaudi-3-ai-accelerator.html", "secondary_source": "" }, { "date": "2024-04-10", "date_label": "", "lab": "meta", "type": "chip", "title": "Meta details its next-generation MTIA chip", "summary": "Meta described a TSMC 5 nm successor to MTIA, already in 16 regions, with a 90 W design and up to three times the performance of the first chip on four models.", "primary_source": "https://ai.meta.com/blog/next-generation-meta-training-inference-accelerator-AI-MTIA/", "secondary_source": "https://techcrunch.com/2024/04/10/meta-unveils-its-newest-custom-ai-chip/" }, { "date": "2024-04-18", "date_label": "", "lab": "meta", "type": "model", "title": "Llama 3 and Meta AI assistant", "summary": "Meta released Llama 3 in 8B and 70B sizes and upgraded its Meta AI assistant.", "primary_source": "https://ai.meta.com/blog/meta-llama-3/", "secondary_source": "https://en.wikipedia.org/wiki/Llama_(language_model)" }, { "date": "2024-05-07", "date_label": "", "lab": "apple", "type": "chip", "title": "Apple M4", "summary": "Apple introduced the M4, with a Neural Engine rated at up to 38 trillion operations per second.", "primary_source": "https://www.apple.com/newsroom/2024/05/apple-introduces-m4-chip/", "secondary_source": "" }, { "date": "2024-05-08", "date_label": "", "lab": "openai", "type": "policy", "title": "Model Spec first published", "summary": "OpenAI published a draft Model Spec describing how its models should behave and resolve conflicts.", "primary_source": "https://openai.com/index/introducing-the-model-spec/", "secondary_source": "https://en.wikipedia.org/wiki/OpenAI" }, { "date": "2024-05-13", "date_label": "", "lab": "openai", "type": "model", "title": "GPT-4o released", "summary": "OpenAI introduced GPT-4o, a natively multimodal model handling text, audio and images, and made it available to free users.", "primary_source": "https://openai.com/index/hello-gpt-4o/", "secondary_source": "https://en.wikipedia.org/wiki/GPT-4o" }, { "date": "2024-05-14", "date_label": "", "lab": "google", "type": "event", "title": "Google I/O 2024: Veo, Imagen 3, AI Overviews", "summary": "Google announced the Veo video model, Imagen 3, Gemini 1.5 Flash and AI Overviews in US Search.", "primary_source": "https://blog.google/technology/ai/google-io-2024-100-announcements/", "secondary_source": "https://en.wikipedia.org/wiki/Veo_(text-to-video_model)" }, { "date": "2024-05-14", "date_label": "", "lab": "google", "type": "chip", "title": "Google Trillium, sixth-generation TPU", "summary": "Google announced Trillium, its sixth TPU generation, claiming 4.7 times the peak compute per chip of TPU v5e.", "primary_source": "https://cloud.google.com/blog/products/compute/introducing-trillium-6th-gen-tpus", "secondary_source": "" }, { "date": "2024-05-20", "date_label": "", "lab": "microsoft", "type": "product", "title": "Copilot+ PCs", "summary": "Microsoft announced Copilot+ PCs with on-device AI hardware.", "primary_source": "https://blogs.microsoft.com/blog/2024/05/20/introducing-copilot-pcs/", "secondary_source": "https://en.wikipedia.org/wiki/Microsoft_Copilot" }, { "date": "2024-05-30", "date_label": "", "lab": "ualink", "type": "infrastructure", "title": "UALink promoter group formed", "summary": "Industry companies formed the Ultra Accelerator Link promoter group to develop an open link between AI accelerators.", "primary_source": "https://ualinkconsortium.org/news/", "secondary_source": "" }, { "date": "2024-06-01", "date_label": "", "lab": "us-doe", "type": "infrastructure", "title": "Aurora passes one exaflop", "summary": "Argonne's Aurora, built on Intel processors and GPUs, became the second system to break the exascale barrier on the TOP500, at 1.012 exaflops.", "primary_source": "https://top500.org/lists/top500/2024/06/", "secondary_source": "" }, { "date": "2024-06-02", "date_label": "", "lab": "nvidia", "type": "chip", "title": "Nvidia names Rubin as next platform", "summary": "Ahead of Computex, Nvidia revealed for the first time that the Rubin platform, with a new Arm-based CPU called Vera, would succeed Blackwell.", "primary_source": "https://blogs.nvidia.com/blog/computex-2024-jensen-huang/", "secondary_source": "" }, { "date": "2024-06-10", "date_label": "", "lab": "apple", "type": "product", "title": "Apple Intelligence announced", "summary": "At WWDC Apple announced Apple Intelligence, on-device and private-cloud AI features, plus optional ChatGPT integration.", "primary_source": "https://www.apple.com/newsroom/2024/06/introducing-apple-intelligence-for-iphone-ipad-and-mac/", "secondary_source": "https://en.wikipedia.org/wiki/Apple_Intelligence" }, { "date": "2024-06-20", "date_label": "", "lab": "anthropic", "type": "model", "title": "Claude 3.5 Sonnet and Artifacts", "summary": "Anthropic released Claude 3.5 Sonnet and Artifacts, a side panel for generated code and documents.", "primary_source": "https://www.anthropic.com/news/claude-3-5-sonnet", "secondary_source": "https://www.theverge.com/2024/6/20/24181961/anthropic-claude-35-sonnet-model-ai-launch" }, { "date": "2024-06-25", "date_label": "", "lab": "etched", "type": "chip", "title": "Etched announces Sohu, a transformer-only chip", "summary": "The startup Etched said it raised $120 million for Sohu, a chip built only for transformer models.", "primary_source": "https://techcrunch.com/2024/06/25/etched-is-building-an-ai-chip-that-only-runs-transformer-models", "secondary_source": "https://www.theregister.com/2024/06/26/etched_asic_ai/" }, { "date": "2024-07-11", "date_label": "", "lab": "softbank", "type": "chip", "title": "SoftBank buys Graphcore", "summary": "SoftBank acquired British AI chip designer Graphcore, with the price undisclosed.", "primary_source": "https://techcrunch.com/2024/07/11/softbank-acquires-uk-ai-chipmaker-graphcore/", "secondary_source": "https://bnnbloomberg.ca/business/technology/2024/07/11/softbank-buys-struggling-uk-chip-startup-graphcore-in-ai-race" }, { "date": "2024-07-18", "date_label": "", "lab": "openai", "type": "model", "title": "GPT-4o mini", "summary": "OpenAI released GPT-4o mini, a small low-cost model that replaced GPT-3.5 Turbo in ChatGPT.", "primary_source": "https://openai.com/index/gpt-4o-mini-advancing-cost-efficient-intelligence/", "secondary_source": "https://en.wikipedia.org/wiki/GPT-4o" }, { "date": "2024-07-18", "date_label": "", "lab": "tenstorrent", "type": "hardware", "title": "Tenstorrent puts Wormhole cards on sale", "summary": "Tenstorrent opened orders for Wormhole n150s and n300s PCIe cards and workstations aimed at developers.", "primary_source": "https://tenstorrent.com/vision/tenstorrent-launches-next-generation-wormhole-based-developer-kits-and-workstations", "secondary_source": "https://at-web1.www.anandtech.com/show/21482/tenstorrent-launches-wormhole-ai-processors-466-fp8-tflops-at-300w" }, { "date": "2024-07-23", "date_label": "", "lab": "meta", "type": "model", "title": "Llama 3.1 405B", "summary": "Meta released Llama 3.1, including a 405-billion-parameter open-weight model.", "primary_source": "https://ai.meta.com/blog/meta-llama-3-1/", "secondary_source": "https://en.wikipedia.org/wiki/Llama_(language_model)" }, { "date": "2024-07-25", "date_label": "", "lab": "openai", "type": "product", "title": "SearchGPT prototype", "summary": "OpenAI tested SearchGPT, a search prototype with sourced answers, with a small group of users.", "primary_source": "https://openai.com/index/searchgpt-prototype/", "secondary_source": "https://en.wikipedia.org/wiki/ChatGPT_Search" }, { "date": "2024-07-30", "date_label": "", "lab": "midjourney", "type": "model", "title": "Midjourney V6.1", "summary": "Midjourney released V6.1, available on Discord and on its website.", "primary_source": "https://updates.midjourney.com/version-6-1/", "secondary_source": "https://en.wikipedia.org/wiki/Midjourney" }, { "date": "2024-08-06", "date_label": "", "lab": "openai", "type": "api", "title": "Structured Outputs in the API", "summary": "OpenAI added Structured Outputs so model responses reliably match a developer-supplied JSON schema.", "primary_source": "https://openai.com/index/introducing-structured-outputs-in-the-api/", "secondary_source": "https://en.wikipedia.org/wiki/OpenAI" }, { "date": "2024-08-27", "date_label": "", "lab": "cerebras", "type": "hardware", "title": "Cerebras Inference", "summary": "Cerebras launched an inference service, claiming 1,800 tokens per second on Llama 3.1 8B.", "primary_source": "https://www.cerebras.ai/press-release/cerebras-launches-the-worlds-fastest-ai-inference", "secondary_source": "" }, { "date": "2024-09-02", "date_label": "", "lab": "xai", "type": "infrastructure", "title": "xAI brings Colossus online", "summary": "Elon Musk said xAI's Colossus cluster of 100,000 Nvidia GPUs had come online.", "primary_source": "https://siliconangle.com/2024/09/03/elon-musks-xai-launches-colossus-ai-training-system-100000-nvidia-chips/", "secondary_source": "https://x.ai/colossus" }, { "date": "2024-09-06", "date_label": "", "lab": "netherlands", "type": "policy", "title": "Netherlands widens licensing to more ASML tools", "summary": "The Dutch government announced, effective the next day, an expanded national export control on advanced chipmaking equipment.", "primary_source": "https://www.government.nl/latest/news/2024/09/06/the-netherlands-expands-export-control-measure-advanced-semiconductor-manufacturing-equipment", "secondary_source": "https://www.trendforce.com/news/2024/09/09/news-netherlands-expands-export-control-over-asmls-two-duv-machines-effective-on-september-7th/" }, { "date": "2024-09-12", "date_label": "", "lab": "openai", "type": "model", "title": "OpenAI o1-preview and o1-mini", "summary": "OpenAI released o1-preview and o1-mini, reasoning models that think before answering.", "primary_source": "https://openai.com/index/introducing-openai-o1-preview/", "secondary_source": "https://en.wikipedia.org/wiki/OpenAI_o1" }, { "date": "2024-09-14", "date_label": "", "lab": "cscs", "type": "infrastructure", "title": "Alps supercomputer inaugurated", "summary": "ETH Zurich inaugurated the Alps supercomputer at the Swiss National Supercomputing Centre in Lugano.", "primary_source": "https://www.cscs.ch/publications/news/2024/new-research-infrastructure-alps-supercomputer-inaugurated", "secondary_source": "" }, { "date": "2024-09-25", "date_label": "", "lab": "meta", "type": "model", "title": "Llama 3.2 vision and edge models", "summary": "Meta released Llama 3.2 with vision models and small models for phones.", "primary_source": "https://ai.meta.com/blog/llama-3-2-connect-2024-vision-edge-mobile-devices/", "secondary_source": "https://en.wikipedia.org/wiki/Llama_(language_model)" }, { "date": "2024-09-26", "date_label": "", "lab": "sk-hynix", "type": "chip", "title": "SK hynix 12-layer HBM3E volume production", "summary": "SK hynix began volume production of 36 GB, 12-layer HBM3E.", "primary_source": "https://news.skhynix.com/en/sk-hynix-begins-volume-production-of-the-world-first-12-layer-hbm3e/", "secondary_source": "" }, { "date": "2024-10-01", "date_label": "", "lab": "openai", "type": "api", "title": "DevDay 2024: Realtime API", "summary": "OpenAI announced the Realtime API for speech-to-speech apps, plus prompt caching, vision fine-tuning and model distillation.", "primary_source": "https://openai.com/index/introducing-the-realtime-api/", "secondary_source": "https://en.wikipedia.org/wiki/OpenAI" }, { "date": "2024-10-03", "date_label": "", "lab": "openai", "type": "feature", "title": "Canvas in ChatGPT", "summary": "OpenAI introduced canvas, a side-by-side workspace for writing and coding with ChatGPT.", "primary_source": "https://openai.com/index/introducing-canvas/", "secondary_source": "https://en.wikipedia.org/wiki/ChatGPT" }, { "date": "2024-10-10", "date_label": "", "lab": "amd", "type": "chip", "title": "AMD Instinct MI325X", "summary": "AMD announced the MI325X accelerator with 256 GB of HBM3E memory.", "primary_source": "https://www.amd.com/en/newsroom/press-releases/2024-10-10-amd-delivers-leadership-ai-performance-with-amd-in.html", "secondary_source": "" }, { "date": "2024-10-22", "date_label": "", "lab": "anthropic", "type": "feature", "title": "Computer use beta, new 3.5 Sonnet and 3.5 Haiku", "summary": "Anthropic let Claude control a computer through screenshots, mouse and keyboard in public beta, alongside upgraded models.", "primary_source": "https://www.anthropic.com/news/3-5-models-and-computer-use", "secondary_source": "https://www.theverge.com/2024/10/22/24276822/anthopic-claude-3-5-sonnet-computer-use-ai" }, { "date": "2024-10-28", "date_label": "", "lab": "apple", "type": "product", "title": "First Apple Intelligence features ship", "summary": "Apple released iOS 18.1 with the first Apple Intelligence features in US English.", "primary_source": "https://www.apple.com/newsroom/2024/10/apple-intelligence-is-available-today-on-iphone-ipad-and-mac/", "secondary_source": "https://en.wikipedia.org/wiki/Apple_Intelligence" }, { "date": "2024-10-29", "date_label": "", "lab": "ualink", "type": "infrastructure", "title": "UALink Consortium incorporates", "summary": "The Ultra Accelerator Link group incorporated and opened membership.", "primary_source": "https://ualinkconsortium.org/news/", "secondary_source": "" }, { "date": "2024-10-31", "date_label": "", "lab": "openai", "type": "feature", "title": "ChatGPT search", "summary": "OpenAI built web search with source links into ChatGPT, starting with paid users.", "primary_source": "https://openai.com/index/introducing-chatgpt-search/", "secondary_source": "https://en.wikipedia.org/wiki/ChatGPT_Search" }, { "date": "2024-11-01", "date_label": "", "lab": "us-doe", "type": "infrastructure", "title": "El Capitan tops the TOP500", "summary": "Lawrence Livermore's El Capitan, on AMD MI300A accelerators, took first place on the TOP500 with 1.742 exaflops.", "primary_source": "https://www.top500.org/news/el-capitan-achieves-top-spot-frontier-and-aurora-follow-behind/", "secondary_source": "" }, { "date": "2024-11-25", "date_label": "", "lab": "anthropic", "type": "api", "title": "Model Context Protocol (MCP)", "summary": "Anthropic open-sourced MCP, a standard for connecting AI assistants to data sources and tools.", "primary_source": "https://www.anthropic.com/news/model-context-protocol", "secondary_source": "https://en.wikipedia.org/wiki/Model_Context_Protocol" }, { "date": "2024-12", "date_label": "", "lab": "meta", "type": "infrastructure", "title": "Meta breaks ground in Richland Parish, Louisiana", "summary": "Meta broke ground on its Richland Parish data center.", "primary_source": "https://datacenters.atmeta.com/2026/07/deepening-our-investment-in-richland-parish-louisiana/", "secondary_source": "https://www.fortune.com/2026/03/27/meta-hyperion-10-gas-power-plants-louisiana-entergy/" }, { "date": "2024-12-02", "date_label": "", "lab": "us-commerce", "type": "policy", "title": "US controls on HBM and new foreign direct product rules", "summary": "The US Commerce Department added controls on High Bandwidth Memory and new foreign direct product rules for chipmaking equipment.", "primary_source": "https://www.federalregister.gov/documents/2024/12/05/2024-28270/foreign-produced-direct-product-rule-additions-and-refinements-to-controls-for-advanced-computing", "secondary_source": "" }, { "date": "2024-12-03", "date_label": "", "lab": "china", "type": "policy", "title": "China bans gallium and germanium exports to the US", "summary": "China's commerce ministry barred exports of gallium, germanium, antimony and superhard materials to the United States.", "primary_source": "https://cset.georgetown.edu/publication/china-rare-earth-export-ban/", "secondary_source": "https://abcnews.go.com/US/wireStory/china-bans-exports-us-gallium-key-materials-after-116399515" }, { "date": "2024-12-03", "date_label": "", "lab": "amazon", "type": "chip", "title": "AWS Trn2 instances generally available", "summary": "AWS made EC2 Trn2 instances generally available.", "primary_source": "https://aws.amazon.com/about-aws/whats-new/2024/12/amazon-ec2-trn2-instances-available/", "secondary_source": "" }, { "date": "2024-12-04", "date_label": "", "lab": "meta", "type": "infrastructure", "title": "Louisiana announces Meta's $10 billion Richland Parish data centre", "summary": "Louisiana's governor announced that Meta chose Richland Parish for a $10 billion AI data centre.", "primary_source": "https://gov.louisiana.gov/news/4697", "secondary_source": "https://www.kplctv.com/2024/12/05/meta-build-10-billion-ai-data-center-richland-parish" }, { "date": "2024-12-05", "date_label": "", "lab": "openai", "type": "product", "title": "Full o1 and ChatGPT Pro", "summary": "OpenAI released the full o1 model and a $200 monthly Pro plan with o1 pro mode.", "primary_source": "https://openai.com/index/introducing-chatgpt-pro/", "secondary_source": "https://en.wikipedia.org/wiki/OpenAI_o1" }, { "date": "2024-12-05", "date_label": "", "lab": "canada", "type": "policy", "title": "Canadian Sovereign AI Compute Strategy", "summary": "Canada launched a strategy investing up to $2 billion in AI computing: data centres, a public supercomputer and help for small businesses to pay for compute.", "primary_source": "https://www.canada.ca/en/innovation-science-economic-development/news/2024/12/canada-to-drive-billions-in-investments-to-build-domestic-ai-compute-capacity-at-home.html", "secondary_source": "https://www.canada.ca/en/department-finance/news/2024/12/deputy-prime-minister-announces-240-million-for-cohere-to-scale-up-ai-compute-capacity.html" }, { "date": "2024-12-09", "date_label": "", "lab": "openai", "type": "product", "title": "Sora released to the public", "summary": "OpenAI launched Sora Turbo video generation for ChatGPT Plus and Pro users at sora.com.", "primary_source": "https://openai.com/index/sora-is-here/", "secondary_source": "https://en.wikipedia.org/wiki/Sora_(text-to-video_model)" }, { "date": "2024-12-09", "date_label": "", "lab": "microsoft", "type": "infrastructure", "title": "Microsoft announces zero-water-for-cooling data center design", "summary": "Microsoft said all new data center designs from August 2024 use closed-loop chip-level cooling that avoids evaporating water.", "primary_source": "https://www.microsoft.com/en-us/microsoft-cloud/blog/2024/12/09/sustainable-by-design-next-generation-datacenters-consume-zero-water-for-cooling/", "secondary_source": "" }, { "date": "2024-12-10", "date_label": "", "lab": "eu", "type": "infrastructure", "title": "EU selects its first seven AI Factories", "summary": "EuroHPC selected seven AI Factories in Finland, Germany, Greece, Italy, Luxembourg, Spain and Sweden.", "primary_source": "https://www.eurohpc-ju.europa.eu/selection-first-seven-ai-factories-drive-europes-leadership-ai-2024-12-10_en", "secondary_source": "" }, { "date": "2024-12-11", "date_label": "", "lab": "google", "type": "model", "title": "Gemini 2.0 Flash experimental", "summary": "Google introduced Gemini 2.0, starting with an experimental Flash model aimed at agentic uses.", "primary_source": "https://blog.google/technology/google-deepmind/google-gemini-ai-update-december-2024/", "secondary_source": "https://en.wikipedia.org/wiki/Gemini_(language_model)" }, { "date": "2024-12-11", "date_label": "", "lab": "google", "type": "product", "title": "Google's Trillium TPU reaches general availability", "summary": "Google made its sixth-generation TPU generally available and said it trained Gemini 2.0 on it.", "primary_source": "https://cloud.google.com/blog/products/compute/trillium-tpu-is-ga", "secondary_source": "" }, { "date": "2024-12-17", "date_label": "", "lab": "openai", "type": "api", "title": "o1 in the API", "summary": "OpenAI brought o1 to the API with function calling and structured outputs, and updated the Realtime API.", "primary_source": "https://openai.com/index/o1-and-new-tools-for-developers/", "secondary_source": "https://en.wikipedia.org/wiki/OpenAI_o1" }, { "date": "2024-12-20", "date_label": "", "lab": "openai", "type": "model", "title": "o3 and o3-mini previewed", "summary": "On the last day of its '12 days' event OpenAI previewed o3 and o3-mini for safety testing, not public release.", "primary_source": "https://openai.com/index/early-access-for-safety-testing/", "secondary_source": "https://en.wikipedia.org/wiki/OpenAI_o3" }, { "date": "2024-12-26", "date_label": "", "lab": "deepseek", "type": "model", "title": "DeepSeek-V3", "summary": "DeepSeek released V3, a 671B mixture-of-experts open model reportedly trained at low cost.", "primary_source": "https://arxiv.org/abs/2412.19437", "secondary_source": "https://en.wikipedia.org/wiki/DeepSeek" }, { "date": "2025-01-06", "date_label": "", "lab": "nvidia", "type": "chip", "title": "GeForce RTX 50 series (Blackwell)", "summary": "At CES, Nvidia announced the Blackwell-based GeForce RTX 50 series, led by the RTX 5090 with 92 billion transistors.", "primary_source": "https://nvidianews.nvidia.com/news/nvidia-blackwell-geforce-rtx-50-series-opens-new-world-of-ai-computer-graphics", "secondary_source": "" }, { "date": "2025-01-06", "date_label": "", "lab": "nvidia", "type": "hardware", "title": "Nvidia Project Digits desktop", "summary": "Nvidia announced Project Digits, a desktop AI computer built on Grace Blackwell, priced from 3,000 US dollars.", "primary_source": "https://nvidianews.nvidia.com/news/nvidia-puts-grace-blackwell-on-every-desk-and-at-every-ai-developers-fingertips", "secondary_source": "" }, { "date": "2025-01-06", "date_label": "", "lab": "amd", "type": "chip", "title": "AMD Ryzen AI Max (Strix Halo)", "summary": "At CES, AMD announced Ryzen AI Max processors with up to 128 GB of unified memory, of which up to 96 GB can serve graphics.", "primary_source": "https://newsroom.amd.com/news/amd-announces-expanded-consumer-and-commercial-ai/", "secondary_source": "" }, { "date": "2025-01-15", "date_label": "", "lab": "us-commerce", "type": "policy", "title": "US tightens foundry checks on AI chips", "summary": "The US issued rules requiring chip factories to do more checks on customers, to stop advanced chips reaching China.", "primary_source": "https://www.bis.gov/press-release/commerce-strengthens-restrictions-advanced-computing-semiconductors-enhance-foundry-due-diligence-prevent", "secondary_source": "https://www.federalregister.gov/documents/2025/01/16/2025-00711/implementation-of-additional-due-diligence-measures-for-advanced-computing-integrated-circuits" }, { "date": "2025-01-15", "date_label": "", "lab": "netherlands", "type": "policy", "title": "Netherlands to license chip measuring tools", "summary": "The Dutch government said that from April 1, 2025, more technology would need a licence, such as specific measuring and inspection equipment.", "primary_source": "https://www.government.nl/latest/news/2025/01/15/klever-export-controls-on-advanced-semiconductor-manufacturing-equipment-to-be-tightened", "secondary_source": "https://www.rijksoverheid.nl/actueel/nieuws/2025/01/15/klever-aanscherping-exportcontrole-op-geavanceerde-productieapparatuur-voor-halfgeleiders" }, { "date": "2025-01-15", "date_label": "", "lab": "us-commerce", "type": "policy", "title": "US AI Diffusion rule", "summary": "The US published the Framework for Artificial Intelligence Diffusion, setting global licence rules for advanced AI chips and some model weights.", "primary_source": "https://www.federalregister.gov/documents/2025/01/15/2025-00636/framework-for-artificial-intelligence-diffusion", "secondary_source": "" }, { "date": "2025-01-20", "date_label": "", "lab": "deepseek", "type": "model", "title": "DeepSeek-R1", "summary": "DeepSeek released R1, an open-weight reasoning model rivaling o1, under the MIT licence.", "primary_source": "https://api-docs.deepseek.com/news/news250120", "secondary_source": "https://en.wikipedia.org/wiki/DeepSeek" }, { "date": "2025-01-21", "date_label": "", "lab": "openai", "type": "event", "title": "Stargate Project announced", "summary": "OpenAI, SoftBank, Oracle and MGX announced a joint venture to build AI data centers in the US.", "primary_source": "https://openai.com/index/announcing-the-stargate-project/", "secondary_source": "https://en.wikipedia.org/wiki/Stargate_LLC" }, { "date": "2025-01-23", "date_label": "", "lab": "openai", "type": "product", "title": "Operator research preview", "summary": "OpenAI released Operator, an agent that uses its own browser to complete web tasks, for US Pro users.", "primary_source": "https://openai.com/index/introducing-operator/", "secondary_source": "https://en.wikipedia.org/wiki/ChatGPT" }, { "date": "2025-01-27", "date_label": "", "lab": "deepseek", "type": "event", "title": "DeepSeek app tops charts; Nvidia shares fall", "summary": "DeepSeek's app reached the top of the US App Store and AI chip stocks dropped sharply.", "primary_source": "", "secondary_source": "https://en.wikipedia.org/wiki/DeepSeek" }, { "date": "2025-01-29", "date_label": "", "lab": "alibaba", "type": "model", "title": "Qwen2.5-Max", "summary": "Alibaba released Qwen2.5-Max, a large mixture-of-experts model.", "primary_source": "https://qwenlm.github.io/blog/qwen2.5-max/", "secondary_source": "https://en.wikipedia.org/wiki/Qwen" }, { "date": "2025-01-31", "date_label": "", "lab": "openai", "type": "model", "title": "OpenAI o3-mini", "summary": "OpenAI released o3-mini, a low-cost reasoning model in ChatGPT (including free users) and the API.", "primary_source": "https://openai.com/index/openai-o3-mini/", "secondary_source": "https://en.wikipedia.org/wiki/OpenAI_o3" }, { "date": "2025-02-02", "date_label": "", "lab": "openai", "type": "product", "title": "Deep research in ChatGPT", "summary": "OpenAI launched deep research, an agent powered by a version of o3 that compiles cited reports from many web sources, for Pro users first.", "primary_source": "https://openai.com/index/introducing-deep-research/", "secondary_source": "https://en.wikipedia.org/wiki/ChatGPT" }, { "date": "2025-02-17", "date_label": "", "lab": "xai", "type": "model", "title": "Grok 3", "summary": "xAI released Grok 3 with a Think reasoning mode.", "primary_source": "https://x.ai/news/grok-3", "secondary_source": "https://en.wikipedia.org/wiki/Grok_(chatbot)" }, { "date": "2025-02-24", "date_label": "", "lab": "anthropic", "type": "model", "title": "Claude 3.7 Sonnet and Claude Code preview", "summary": "Anthropic released Claude 3.7 Sonnet with optional extended thinking and previewed Claude Code, a terminal coding agent.", "primary_source": "https://www.anthropic.com/news/claude-3-7-sonnet", "secondary_source": "https://techcrunch.com/2025/02/24/anthropic-launches-a-new-ai-model-that-thinks-as-long-as-you-want/" }, { "date": "2025-02-27", "date_label": "", "lab": "openai", "type": "model", "title": "GPT-4.5 research preview", "summary": "OpenAI released GPT-4.5, its largest non-reasoning chat model, to Pro users and developers.", "primary_source": "https://openai.com/index/introducing-gpt-4-5/", "secondary_source": "https://en.wikipedia.org/wiki/GPT-4.5" }, { "date": "2025-03-04", "date_label": "", "lab": "tsmc", "type": "infrastructure", "title": "TSMC raises US investment to 165 billion dollars", "summary": "TSMC said it intended to expand its total United States investment to 165 billion US dollars.", "primary_source": "https://pr.tsmc.com/english/news/3210", "secondary_source": "" }, { "date": "2025-03-05", "date_label": "", "lab": "apple", "type": "chip", "title": "Apple M3 Ultra with 512 GB", "summary": "Apple revealed the M3 Ultra, configurable with up to 512 GB of unified memory.", "primary_source": "https://www.apple.com/newsroom/2025/03/apple-reveals-m3-ultra-taking-apple-silicon-to-a-new-extreme/", "secondary_source": "" }, { "date": "2025-03-11", "date_label": "", "lab": "openai", "type": "api", "title": "Responses API and Agents SDK", "summary": "OpenAI launched the Responses API, built-in web search, file search and computer use tools, and the open-source Agents SDK.", "primary_source": "https://openai.com/index/new-tools-for-building-agents/", "secondary_source": "https://en.wikipedia.org/wiki/OpenAI" }, { "date": "2025-03-18", "date_label": "", "lab": "nvidia", "type": "chip", "title": "Nvidia Blackwell Ultra", "summary": "Nvidia announced Blackwell Ultra, including the GB300 NVL72 rack and the HGX B300 NVL16 system.", "primary_source": "https://nvidianews.nvidia.com/news/nvidia-blackwell-ultra-ai-factory-platform-paves-way-for-age-of-ai-reasoning", "secondary_source": "" }, { "date": "2025-03-18", "date_label": "", "lab": "nvidia", "type": "hardware", "title": "Nvidia DGX Spark and DGX Station announced", "summary": "Nvidia announced DGX Spark, the new name for Project Digits, and DGX Station as personal AI computers.", "primary_source": "https://nvidianews.nvidia.com/news/nvidia-announces-dgx-spark-and-dgx-station-personal-ai-computers", "secondary_source": "" }, { "date": "2025-03-18", "date_label": "", "lab": "nvidia", "type": "chip", "title": "Nvidia RTX PRO Blackwell workstation cards", "summary": "Nvidia brought Blackwell to workstations and servers with its RTX PRO line.", "primary_source": "https://nvidianews.nvidia.com/news/nvidia-blackwell-rtx-pro-workstations-servers-agentic-ai", "secondary_source": "" }, { "date": "2025-03-18", "date_label": "", "lab": "nvidia", "type": "infrastructure", "title": "Nvidia co-packaged optics switches", "summary": "Nvidia announced Spectrum-X and Quantum-X Photonics networking switches using co-packaged optics.", "primary_source": "https://nvidianews.nvidia.com/news/nvidia-spectrum-x-co-packaged-optics-networking-switches-ai-factories", "secondary_source": "" }, { "date": "2025-03-20", "date_label": "", "lab": "softbank", "type": "chip", "title": "SoftBank to buy Ampere Computing", "summary": "SoftBank agreed to buy chip designer Ampere Computing in an all-cash deal valued at 6.5 billion US dollars.", "primary_source": "https://group.softbank/en/news/press/20250320", "secondary_source": "" }, { "date": "2025-03-25", "date_label": "", "lab": "openai", "type": "feature", "title": "Native image generation in GPT-4o", "summary": "ChatGPT replaced DALL-E with GPT-4o's built-in image generation, which handles text in images and edits well.", "primary_source": "https://openai.com/index/introducing-4o-image-generation/", "secondary_source": "https://en.wikipedia.org/wiki/GPT-4o" }, { "date": "2025-03-25", "date_label": "", "lab": "google", "type": "model", "title": "Gemini 2.5 Pro", "summary": "Google released Gemini 2.5 Pro, a thinking model that topped several leaderboards.", "primary_source": "https://blog.google/technology/google-deepmind/gemini-model-thinking-updates-march-2025/", "secondary_source": "https://en.wikipedia.org/wiki/Gemini_(language_model)" }, { "date": "2025-03-25", "date_label": "", "lab": "deepseek", "type": "model", "title": "DeepSeek-V3-0324", "summary": "DeepSeek released an updated V3 model.", "primary_source": "https://api-docs.deepseek.com/news/news260910", "secondary_source": "https://huggingface.co/deepseek-ai/DeepSeek-V3-0324" }, { "date": "2025-03-27", "date_label": "", "lab": "coreweave", "type": "finance", "title": "CoreWeave prices its IPO at $40 a share", "summary": "CoreWeave priced its IPO at $40 a share for 37.5 million shares, about $1.5 billion.", "primary_source": "https://investors.coreweave.com/news/news-details/2025/CoreWeave-Announces-Pricing-of-Initial-Public-Offering/default.aspx", "secondary_source": "https://www.cnbc.com/2025/03/27/coreweave-prices-ipo-at-40-a-share-below-expected-range.html" }, { "date": "2025-04-04", "date_label": "", "lab": "midjourney", "type": "model", "title": "Midjourney V7", "summary": "Midjourney released V7 in alpha.", "primary_source": "", "secondary_source": "https://en.wikipedia.org/wiki/Midjourney" }, { "date": "2025-04-05", "date_label": "", "lab": "meta", "type": "model", "title": "Llama 4 Scout and Maverick", "summary": "Meta released Llama 4 Scout and Maverick, mixture-of-experts multimodal models, and previewed Behemoth.", "primary_source": "https://ai.meta.com/blog/llama-4-multimodal-intelligence/", "secondary_source": "https://en.wikipedia.org/wiki/Llama_(language_model)" }, { "date": "2025-04-08", "date_label": "", "lab": "ualink", "type": "infrastructure", "title": "UALink 200G 1.0 specification", "summary": "The UALink Consortium released the Ultra Accelerator Link 200G 1.0 specification.", "primary_source": "https://ualinkconsortium.org/news/", "secondary_source": "" }, { "date": "2025-04-09", "date_label": "", "lab": "google", "type": "chip", "title": "Google Ironwood, seventh-generation TPU", "summary": "Google announced Ironwood, its first TPU built for inference, scaling to 9,216 chips and a claimed 42.5 exaflops.", "primary_source": "https://blog.google/innovation-and-ai/infrastructure-and-cloud/google-cloud/ironwood-tpu-age-of-inference/", "secondary_source": "" }, { "date": "2025-04-14", "date_label": "", "lab": "openai", "type": "model", "title": "GPT-4.1, 4.1 mini and 4.1 nano in the API", "summary": "OpenAI released three API models with up to one million tokens of context and better coding and instruction following.", "primary_source": "https://openai.com/index/gpt-4-1/", "secondary_source": "https://en.wikipedia.org/wiki/GPT-4.1" }, { "date": "2025-04-14", "date_label": "", "lab": "openai", "type": "guide", "title": "GPT-4.1 prompting guide", "summary": "OpenAI published a cookbook guide explaining how to prompt GPT-4.1, which follows instructions more literally.", "primary_source": "https://cookbook.openai.com/examples/gpt4-1_prompting_guide", "secondary_source": "https://openai.com/index/gpt-4-1/" }, { "date": "2025-04-14", "date_label": "", "lab": "nvidia", "type": "policy", "title": "US requires licence for Nvidia H20 to China", "summary": "Nvidia disclosed that a licence requirement for H20 exports to China would stay in effect indefinitely, and expected charges of up to about 5.5 billion US dollars.", "primary_source": "https://www.sec.gov/Archives/edgar/data/1045810/000104581025000082/nvda-20250409.htm", "secondary_source": "" }, { "date": "2025-04-15", "date_label": "", "lab": "telus", "type": "infrastructure", "title": "Telus plans sovereign AI factories", "summary": "Telus announced plans for AI factories in Rimouski and Kamloops.", "primary_source": "https://www.newswire.ca/news-releases/telus-to-launch-sovereign-ai-factories-in-kamloops-and-rimouski-propelling-canada-to-the-forefront-of-ai-innovation-821179363.html", "secondary_source": "" }, { "date": "2025-04-16", "date_label": "", "lab": "openai", "type": "product", "title": "Codex CLI", "summary": "OpenAI released Codex CLI, an open-source coding agent that runs in the terminal.", "primary_source": "https://openai.com/index/introducing-o3-and-o4-mini/", "secondary_source": "https://community.openai.com/t/this-weeks-launches-o3-o4-mini-gpt-4-1-and-codex-cli/1230312" }, { "date": "2025-04-16", "date_label": "", "lab": "openai", "type": "model", "title": "OpenAI o3 and o4-mini", "summary": "OpenAI released o3 and o4-mini, reasoning models that can use all ChatGPT tools and reason with images.", "primary_source": "https://openai.com/index/introducing-o3-and-o4-mini/", "secondary_source": "https://en.wikipedia.org/wiki/OpenAI_o3" }, { "date": "2025-04-16", "date_label": "", "lab": "jedec", "type": "chip", "title": "JEDEC releases the HBM4 standard", "summary": "JEDEC released JESD270-4, the HBM4 standard.", "primary_source": "https://www.jedec.org/news/pressreleases/jedec%C2%AE-and-industry-leaders-collaborate-release-jesd270-4-hbm4-standard-advancing", "secondary_source": "" }, { "date": "2025-04-28", "date_label": "", "lab": "alibaba", "type": "model", "title": "Qwen3", "summary": "Alibaba released the Qwen3 family of open-weight models with hybrid thinking.", "primary_source": "https://qwen.ai/blog?id=qwen3", "secondary_source": "https://en.wikipedia.org/wiki/Qwen" }, { "date": "2025-04-29", "date_label": "", "lab": "openai", "type": "event", "title": "GPT-4o update rolled back for sycophancy", "summary": "OpenAI reverted a GPT-4o update that made ChatGPT overly flattering and explained what went wrong.", "primary_source": "https://openai.com/index/sycophancy-in-gpt-4o/", "secondary_source": "https://en.wikipedia.org/wiki/GPT-4o" }, { "date": "2025-04-30", "date_label": "", "lab": "xiaomi", "type": "model", "title": "Xiaomi MiMo-7B", "summary": "Xiaomi released MiMo-7B, a 7 billion parameter reasoning model, under MIT.", "primary_source": "https://huggingface.co/XiaomiMiMo/MiMo-7B-RL", "secondary_source": "https://en.wikipedia.org/wiki/Xiaomi_MiMo" }, { "date": "2025-05-07", "date_label": "", "lab": "mistral", "type": "model", "title": "Mistral Medium 3", "summary": "Mistral released Mistral Medium 3, a mid-size model aimed at enterprises.", "primary_source": "https://mistral.ai/news/mistral-medium-3/", "secondary_source": "https://en.wikipedia.org/wiki/Mistral_AI" }, { "date": "2025-05-12", "date_label": "", "lab": "humain", "type": "event", "title": "Saudi Arabia launches Humain", "summary": "Saudi Arabia's Crown Prince launched Humain, a Public Investment Fund company spanning the AI value chain.", "primary_source": "https://www.middleeastainews.com/p/saudi-arabia-humain-ai-powerhouse", "secondary_source": "https://www.arabnews.com/business/from-launch-to-ipo-how-saudi-ai-firm-humain-grew-in-16-months-3000770" }, { "date": "2025-05-13", "date_label": "", "lab": "us-bis", "type": "policy", "title": "US Commerce announces rescission of the AI Diffusion rule", "summary": "Commerce said it would rescind the January 2025 three-tier AI Diffusion rule and issued guidance on advanced chips.", "primary_source": "https://www.bis.gov/press-release/department-commerce-announces-rescission-biden-era-artificial-intelligence-diffusion-rule-strengthens", "secondary_source": "https://www.sec.gov/Archives/edgar/data/0001045810/000104581026000075/nvda-20260726.htm" }, { "date": "2025-05-13", "date_label": "", "lab": "us-gov", "type": "policy", "title": "U.S. moves to rescind the AI Diffusion Rule", "summary": "The Commerce Department announced the rescission of the Biden-era AI Diffusion Rule.", "primary_source": "https://www.bis.gov/press-release/department-commerce-announces-rescission-biden-era-artificial-intelligence-diffusion-rule-strengthens", "secondary_source": "" }, { "date": "2025-05-13", "date_label": "", "lab": "us-commerce", "type": "policy", "title": "US says it will rescind AI Diffusion rule", "summary": "The US Commerce Department announced it would rescind the AI Diffusion rule.", "primary_source": "https://www.bis.gov/press-release/department-commerce-announces-rescission-biden-era-artificial-intelligence-diffusion-rule-strengthens", "secondary_source": "" }, { "date": "2025-05-16", "date_label": "", "lab": "openai", "type": "product", "title": "Codex cloud coding agent", "summary": "OpenAI previewed Codex, a cloud software engineering agent powered by codex-1, a version of o3, for Pro, Team and Enterprise users.", "primary_source": "https://openai.com/index/introducing-codex/", "secondary_source": "https://en.wikipedia.org/wiki/OpenAI_Codex" }, { "date": "2025-05-18", "date_label": "", "lab": "nvidia", "type": "infrastructure", "title": "Nvidia NVLink Fusion", "summary": "Nvidia opened NVLink to partners with NVLink Fusion, a way for other companies' custom chips to join its rack-scale systems.", "primary_source": "https://nvidianews.nvidia.com/news/nvidia-nvlink-fusion-semi-custom-ai-infrastructure-partner-ecosystem", "secondary_source": "" }, { "date": "2025-05-19", "date_label": "", "lab": "intel", "type": "chip", "title": "Intel launches the Arc Pro B60 with 24 GB", "summary": "Intel launched Arc Pro B60 and B50 workstation GPUs, with up to 24 GB per card and AI inference as a stated use.", "primary_source": "https://www.businesswire.com/news/home/20250519565090/en/Computex-2025-Intel-Unveils-New-GPUs-for-AI-and-Workstations", "secondary_source": "https://hothardware.com/news/intel-unveils-arc-pro-b60-and-b50-gpus" }, { "date": "2025-05-22", "date_label": "", "lab": "anthropic", "type": "model", "title": "Claude Opus 4 and Sonnet 4", "summary": "Anthropic released Claude Opus 4 and Sonnet 4 and made Claude Code generally available.", "primary_source": "https://www.anthropic.com/news/claude-4", "secondary_source": "https://www.cnbc.com/2025/05/22/claude-4-opus-sonnet-anthropic.html" }, { "date": "2025-05-22", "date_label": "", "lab": "g42", "type": "infrastructure", "title": "Stargate UAE announced", "summary": "G42, OpenAI, Oracle, Nvidia, SoftBank Group and Cisco announced a 1-gigawatt cluster within a 5-gigawatt UAE-U.S. campus.", "primary_source": "https://www.g42.ai/resources/news/global-tech-alliance-launches-stargate-uae", "secondary_source": "https://www.cnbc.com/2026/09/15/aws-cant-restore-service-to-bahrain-uae-6-months-after-iran-strikes.html" }, { "date": "2025-05-28", "date_label": "", "lab": "deepseek", "type": "model", "title": "DeepSeek-R1-0528", "summary": "DeepSeek released an updated R1.", "primary_source": "https://api-docs.deepseek.com/news/news260910", "secondary_source": "https://huggingface.co/deepseek-ai/DeepSeek-R1-0528" }, { "date": "2025-05-28", "date_label": "", "lab": "bell", "type": "infrastructure", "title": "Bell launches AI Fabric", "summary": "Bell announced a plan for upwards of 500 MW of AI compute across six British Columbia facilities.", "primary_source": "https://explore.business.bell.ca/news-and-events/increasing-sovereign-ai-capacity-introducing-bell-ai-fabric", "secondary_source": "" }, { "date": "2025-05-29", "date_label": "", "lab": "us-doe", "type": "hardware", "title": "US names the Doudna supercomputer", "summary": "The US energy secretary announced a Dell contract for NERSC-10, named Doudna and due in 2026, using Nvidia technology.", "primary_source": "https://www.energy.gov/articles/doe-announces-new-supercomputer-powered-dell-and-nvidia-speed-scientific-discovery", "secondary_source": "https://newscenter.lbl.gov/2025/05/29/doe-announces-new-supercomputer-powered-by-dell-and-nvidia-to-speed-scientific-discovery/" }, { "date": "2025-06", "date_label": "", "lab": "minimax", "type": "model", "title": "MiniMax-M1", "summary": "MiniMax released M1 under Apache 2.0 with about 456 billion parameters.", "primary_source": "https://huggingface.co/MiniMaxAI/MiniMax-M1-80k", "secondary_source": "https://en.wikipedia.org/wiki/MiniMax_(company)" }, { "date": "2025-06-01", "date_label": "", "lab": "openai", "type": "infrastructure", "title": "First GB200 racks reach Stargate Abilene", "summary": "OpenAI said Oracle began delivering the first Nvidia GB200 racks to Abilene in June 2025.", "primary_source": "https://openai.com/index/stargate-advances-with-partnership-with-oracle/", "secondary_source": "https://www.rcrwireless.com/20251001/uncategorized/crusoe-ai" }, { "date": "2025-06-03", "date_label": "", "lab": "broadcom", "type": "chip", "title": "Broadcom ships Tomahawk 6 at 102.4 Tb/s", "summary": "Broadcom began shipping a switch series with 102.4 Tb/s of capacity, which it called the world's first.", "primary_source": "https://investors.broadcom.com/news-releases/news-release-details/broadcom-ships-tomahawk-6-worlds-first-1024-tbps-switch", "secondary_source": "https://www.storagereview.com/news/broadcom-ships-tomahawk-6-switch-series-unleashing-102-4-tbps-switching-capacity" }, { "date": "2025-06-05", "date_label": "", "lab": "eleutherai", "type": "paper", "title": "The Common Pile v0.1 and Comma v0.1 models", "summary": "Researchers release an 8 TB dataset of public domain and openly licensed text, with two 7-billion-parameter models trained on it.", "primary_source": "https://arxiv.org/abs/2506.05209", "secondary_source": "https://blog.eleuther.ai/common-pile/" }, { "date": "2025-06-10", "date_label": "", "lab": "mistral", "type": "model", "title": "Magistral reasoning models", "summary": "Mistral released Magistral, its first reasoning models.", "primary_source": "https://mistral.ai/news/magistral", "secondary_source": "https://en.wikipedia.org/wiki/Mistral_AI" }, { "date": "2025-06-11", "date_label": "", "lab": "uec", "type": "infrastructure", "title": "Ultra Ethernet 1.0 specification", "summary": "The Ultra Ethernet Consortium published its 1.0 specification, an Ethernet-based stack for AI and high-performance computing.", "primary_source": "https://www.linuxfoundation.org/press/uec-launches-spec-1.0", "secondary_source": "" }, { "date": "2025-06-12", "date_label": "", "lab": "amd", "type": "chip", "title": "AMD Instinct MI350 series", "summary": "At Advancing AI 2025, AMD launched the MI350 series and laid out a rack-scale roadmap.", "primary_source": "https://newsroom.amd.com/news/amd-unveils-vision-for-an-open-ai-ecosystem-detai/", "secondary_source": "" }, { "date": "2025-06-15", "date_label": "", "lab": "taiwan", "type": "policy", "title": "Taiwan puts Huawei and SMIC on its export control list", "summary": "Taiwan's Commerce Ministry added Huawei and SMIC to its export control list of strategic high-tech commodities.", "primary_source": "https://mynews13.com/fl/orlando/ap-top-news/2025/06/15/taiwan-adds-chinas-huawei-and-smic-to-export-control-list", "secondary_source": "https://www.gigazine.net/gsc_news/en/20250616-taiwan-bans-chip-exports-to-huawei-smic" }, { "date": "2025-06-18", "date_label": "", "lab": "midjourney", "type": "model", "title": "Midjourney V1 video model", "summary": "Midjourney released its first video model, animating images into short clips.", "primary_source": "https://updates.midjourney.com/introducing-our-v1-video-model/", "secondary_source": "https://techcrunch.com/2025/06/18/midjourney-launches-its-first-ai-video-generation-model-v1/" }, { "date": "2025-06-18", "date_label": "", "lab": "calcul-quebec", "type": "hardware", "title": "Calcul Québec launches Rorqual", "summary": "Calcul Québec launched Rorqual, a research supercomputer at the McGill data centre, ranked 192nd on the June 2025 TOP500 for its GPU part.", "primary_source": "https://www.calculquebec.ca/en/?p=6871", "secondary_source": "https://www.alliancecan.ca/en/latest/news/rorqual-supercomputer-welcomed-research-community" }, { "date": "2025-07", "date_label": "", "lab": "zai", "type": "model", "title": "GLM-4.5", "summary": "Z.ai released GLM-4.5 under MIT, with 355 billion total parameters by its model card.", "primary_source": "https://huggingface.co/zai-org/GLM-4.5", "secondary_source": "https://en.wikipedia.org/wiki/Z.ai" }, { "date": "2025-07-08", "date_label": "", "lab": "huggingface", "type": "model", "title": "SmolLM3", "summary": "Hugging Face publishes a 3-billion-parameter open model trained on about 11 trillion tokens with 384 H100 GPUs over 24 days, and details its training recipe.", "primary_source": "https://huggingface.co/blog/smollm3", "secondary_source": "" }, { "date": "2025-07-09", "date_label": "", "lab": "xai", "type": "model", "title": "Grok 4 and Grok 4 Heavy", "summary": "xAI released Grok 4 and a multi-agent Grok 4 Heavy tier.", "primary_source": "https://x.ai/news/grok-4", "secondary_source": "https://en.wikipedia.org/wiki/Grok_(chatbot)" }, { "date": "2025-07-11", "date_label": "", "lab": "moonshot", "type": "model", "title": "Kimi K2 open weights", "summary": "Moonshot released Kimi K2, a mixture-of-experts model with 1 trillion parameters and 32 billion active, under a Modified MIT licence.", "primary_source": "https://huggingface.co/moonshotai/Kimi-K2-Instruct", "secondary_source": "https://en.wikipedia.org/wiki/Moonshot_AI" }, { "date": "2025-07-14", "date_label": "", "lab": "meta", "type": "infrastructure", "title": "Zuckerberg names Meta's Prometheus and Hyperion clusters", "summary": "Mark Zuckerberg said Meta was building Prometheus, due online in 2026, and Hyperion, able to scale to 5 gigawatts over several years.", "primary_source": "https://www.threads.com/@zuck/post/DMF6uUgx9f9", "secondary_source": "https://techcrunch.com/2025/07/14/mark-zuckerberg-says-meta-is-building-a-5gw-ai-data-center" }, { "date": "2025-07-14", "date_label": "", "lab": "malaysia", "type": "policy", "title": "Malaysia requires permits for US AI chips", "summary": "Malaysia's trade ministry said exports, transhipments and transits of high-performance US-origin AI chips needed a Strategic Trade Permit, effective at once.", "primary_source": "https://www.miti.gov.my/miti/resources/Media%20Release/[FINAL]%5FMITI%5FPress%5FStmt%5FMalaysia%5FRegulates%5FTrade%5Fof%5FUS%5FAI%5FChips%5F2025-07-14.pdf", "secondary_source": "https://www.malaymail.com/news/malaysia/2025/07/14/malaysia-acts-to-avoid-us-sanctions-amid-escalating-global-ai-chip-war/183871" }, { "date": "2025-07-14", "date_label": "", "lab": "nvidia", "type": "policy", "title": "Nvidia says H20 sales to China will resume", "summary": "Nvidia said it was filing for licences to sell the H20 again and that the US government had assured it licences would be granted.", "primary_source": "https://blogs.nvidia.com/blog/nvidia-ceo-promotes-ai-in-dc-and-china/", "secondary_source": "" }, { "date": "2025-07-15", "date_label": "", "lab": "broadcom", "type": "chip", "title": "Broadcom ships Tomahawk Ultra", "summary": "Broadcom shipped an Ethernet switch aimed at AI scale-up, with 250 ns latency at 51.2 Tb/s.", "primary_source": "https://investors.broadcom.com/news-releases/news-release-details/broadcom-ships-tomahawk-ultra-reimagining-ethernet-switch-hpc", "secondary_source": "" }, { "date": "2025-07-17", "date_label": "", "lab": "openai", "type": "product", "title": "ChatGPT agent", "summary": "OpenAI merged Operator and deep research into ChatGPT agent, which uses its own virtual computer to complete tasks.", "primary_source": "https://openai.com/index/introducing-chatgpt-agent/", "secondary_source": "https://en.wikipedia.org/wiki/ChatGPT" }, { "date": "2025-07-17", "date_label": "", "lab": "uk-gov", "type": "infrastructure", "title": "UK launches Isambard-AI", "summary": "The UK's most powerful supercomputer, Isambard-AI, launched in Bristol.", "primary_source": "https://www.bristol.ac.uk/news/2025/july/isambard-launch.html", "secondary_source": "" }, { "date": "2025-07-22", "date_label": "", "lab": "openai", "type": "infrastructure", "title": "OpenAI and Oracle add 4.5 GW to Stargate", "summary": "OpenAI and Oracle agreed to develop 4.5 gigawatts of additional Stargate capacity in the United States.", "primary_source": "https://openai.com/index/stargate-advances-with-partnership-with-oracle/", "secondary_source": "" }, { "date": "2025-07-23", "date_label": "", "lab": "us-gov", "type": "policy", "title": "U.S. order on exporting the American AI stack", "summary": "A U.S. executive order set up a program to promote full-stack American AI export packages.", "primary_source": "https://www.whitehouse.gov/presidential-actions/2025/07/promoting-the-export-of-the-american-ai-technology-stack/", "secondary_source": "" }, { "date": "2025-08-04", "date_label": "", "lab": "broadcom", "type": "chip", "title": "Broadcom ships Jericho4", "summary": "Broadcom shipped a router designed to link over one million accelerators across several data centres.", "primary_source": "https://investors.broadcom.com/news-releases/news-release-details/broadcom-ships-jericho4-enabling-distributed-ai-computing-across", "secondary_source": "" }, { "date": "2025-08-05", "date_label": "", "lab": "openai", "type": "model", "title": "gpt-oss open-weight models", "summary": "OpenAI released gpt-oss-120b and gpt-oss-20b, its first open-weight language models since GPT-2, under Apache 2.0.", "primary_source": "https://openai.com/index/introducing-gpt-oss/", "secondary_source": "https://lifearchitect.substack.com/p/the-memo-special-edition-openai-gpt" }, { "date": "2025-08-05", "date_label": "", "lab": "anthropic", "type": "model", "title": "Claude Opus 4.1", "summary": "Anthropic released Opus 4.1 with improved coding and agentic performance.", "primary_source": "https://www.anthropic.com/news/claude-opus-4-1", "secondary_source": "https://en.wikipedia.org/wiki/Claude_(language_model)" }, { "date": "2025-08-07", "date_label": "", "lab": "openai", "type": "model", "title": "GPT-5 released", "summary": "OpenAI launched GPT-5, a unified system that routes between fast and thinking models, to all ChatGPT users including free.", "primary_source": "https://openai.com/index/introducing-gpt-5/", "secondary_source": "https://www.theverge.com/openai/748017/gpt-5-chatgpt-openai-release" }, { "date": "2025-08-07", "date_label": "", "lab": "scinet", "type": "hardware", "title": "SciNet launches the Trillium supercomputer in Toronto", "summary": "The University of Toronto's SciNet launched Trillium, with 1,224 CPU nodes and a GPU section, replacing Niagara and Mist.", "primary_source": "https://scinethpc.ca/trillium/", "secondary_source": "https://betakit.com/inside-the-gta-facility-housing-one-of-canadas-most-powerful-supercomputers/" }, { "date": "2025-08-07", "date_label": "", "lab": "tesla", "type": "chip", "title": "Tesla disbands Dojo team", "summary": "Bloomberg reported that Tesla was disbanding its Dojo supercomputer team and that its leader, Peter Bannon, was leaving.", "primary_source": "https://www.ttnews.com/articles/tesla-disbands-dojo", "secondary_source": "https://the-decoder.com/tesla-has-shut-down-its-dojo-supercomputer-project-and-reassigned-the-entire-team/" }, { "date": "2025-08-11", "date_label": "", "lab": "zai", "type": "model", "title": "GLM-4.5V", "summary": "Z.ai released GLM-4.5V, a vision-language model of 108 billion parameters, under MIT.", "primary_source": "https://huggingface.co/zai-org/GLM-4.5V", "secondary_source": "https://en.wikipedia.org/wiki/Z.ai" }, { "date": "2025-08-21", "date_label": "", "lab": "deepseek", "type": "model", "title": "DeepSeek-V3.1", "summary": "DeepSeek released V3.1 with hybrid thinking and non-thinking modes.", "primary_source": "https://api-docs.deepseek.com/news/news250821", "secondary_source": "https://en.wikipedia.org/wiki/DeepSeek" }, { "date": "2025-08-22", "date_label": "", "lab": "us-gov", "type": "finance", "title": "US takes an equity stake in Intel", "summary": "Intel said the US government would invest $8.9 billion in its common stock, funded largely by CHIPS Act grants not yet paid.", "primary_source": "https://newsroom.intel.com/corporate/intel-and-trump-administration-reach-historic-agreement", "secondary_source": "https://www.intc.com/news-events/press-releases/detail/1748/intel-and-trump-administration-reach-historic-agreement-to" }, { "date": "2025-08-26", "date_label": "", "lab": "google", "type": "model", "title": "Gemini 2.5 Flash Image (Nano Banana)", "summary": "Google released an image editing and generation model, nicknamed Nano Banana, in the Gemini app.", "primary_source": "https://blog.google/products/gemini/updated-image-editing-model/", "secondary_source": "https://en.wikipedia.org/wiki/Gemini_(language_model)" }, { "date": "2025-08-28", "date_label": "", "lab": "openai", "type": "api", "title": "gpt-realtime and Realtime API GA", "summary": "OpenAI made the Realtime API generally available and released gpt-realtime, a speech-to-speech model with SIP calling and MCP support.", "primary_source": "https://openai.com/index/introducing-gpt-realtime/", "secondary_source": "https://gigazine.net/news/20250829-openai-gpt-realtime/" }, { "date": "2025-09", "date_label": "", "lab": "zai", "type": "model", "title": "GLM-4.6", "summary": "Z.ai released GLM-4.6 under MIT.", "primary_source": "https://huggingface.co/zai-org/GLM-4.6", "secondary_source": "https://en.wikipedia.org/wiki/Z.ai" }, { "date": "2025-09-02", "date_label": "", "lab": "swiss-ai", "type": "model", "title": "Apertus, Switzerland's public open model", "summary": "EPFL, ETH Zurich and CSCS released Apertus, a fully open multilingual model (8B and 70B) with public weights, data recipes and code, presented as public AI.", "primary_source": "https://ai.epfl.ch/apertus-a-fully-open-transparent-multilingual-language-model/", "secondary_source": "" }, { "date": "2025-09-05", "date_label": "", "lab": "eurohpc", "type": "hardware", "title": "JUPITER inaugurated at Jülich", "summary": "Germany's chancellor and an EU commissioner inaugurated JUPITER, Europe's first exascale machine, with about 24,000 Nvidia GH200 chips.", "primary_source": "https://eurohpc-ju.europa.eu/jupiter-launching-europes-exascale-era-2025-09-05_en", "secondary_source": "https://atos.net/en/2025/press-release_2025_09_05/inauguration-of-jupiter-europes-most-powerful-supercomputer" }, { "date": "2025-09-09", "date_label": "", "lab": "mistral", "type": "finance", "title": "Mistral AI raises €1.7 billion", "summary": "Mistral AI announced a €1.7 billion Series C at an €11.7 billion post-money valuation.", "primary_source": "https://mistral.ai/news/mistral-ai-raises-1-7-b-to-accelerate-technological-progress-with-ai", "secondary_source": "" }, { "date": "2025-09-12", "date_label": "", "lab": "sk-hynix", "type": "chip", "title": "SK hynix completes HBM4 development", "summary": "SK hynix said it had completed HBM4 development and was ready for mass production, with speeds above 10 gigabits per second.", "primary_source": "https://news.skhynix.com/en/sk-hynix-completes-worlds-first-hbm4-development-and-readies-mass-production/", "secondary_source": "" }, { "date": "2025-09-15", "date_label": "", "lab": "openai", "type": "model", "title": "GPT-5-Codex", "summary": "OpenAI released GPT-5-Codex, a GPT-5 version tuned for agentic coding, as the default for Codex cloud tasks and code review.", "primary_source": "https://openai.com/index/introducing-upgrades-to-codex/", "secondary_source": "https://gigazine.net/gsc_news/en/20250916-openai-gpt-5-codex" }, { "date": "2025-09-15", "date_label": "", "lab": "sfu", "type": "hardware", "title": "SFU brings Fir online", "summary": "Simon Fraser University announced Fir, replacing Cedar after about $80 million in upgrades, as Canada's most powerful academic supercomputer.", "primary_source": "https://www.sfu.ca/sfunews/media/media-releases/2025/09/canada-s-fastest-academic-supercomputer-is-now-online-at-sfu-aft.html", "secondary_source": "https://alliancecan.ca/en/latest/news/supercomputer-fir-now-fully-operational" }, { "date": "2025-09-18", "date_label": "", "lab": "huawei", "type": "chip", "title": "Huawei publishes its Ascend roadmap to 2028", "summary": "Eric Xu set out Ascend 950PR for early 2026, 950DT for late 2026, 960 for late 2027 and 970 for late 2028, with Huawei's own high-bandwidth memory.", "primary_source": "https://www.huawei.com/en/news/2025/9/hc-xu-keynote-speech", "secondary_source": "https://www.rcrwireless.com/20250922/ai-infrastructure/huawei-ai-chips" }, { "date": "2025-09-18", "date_label": "", "lab": "nvidia", "type": "chip", "title": "Nvidia and Intel partner on data centre and PC chips", "summary": "Nvidia said it would invest 5 billion US dollars in Intel, and the two agreed to develop custom data centre and PC products together.", "primary_source": "https://nvidianews.nvidia.com/news/nvidia-and-intel-to-develop-ai-infrastructure-and-personal-computing-products", "secondary_source": "" }, { "date": "2025-09-18", "date_label": "", "lab": "microsoft", "type": "infrastructure", "title": "Microsoft Fairwater", "summary": "Microsoft introduced Fairwater in Wisconsin, which it called its largest AI data centre.", "primary_source": "https://blogs.microsoft.com/blog/2025/09/18/inside-the-worlds-most-powerful-ai-datacenter/", "secondary_source": "" }, { "date": "2025-09-22", "date_label": "", "lab": "openai", "type": "finance", "title": "Nvidia and OpenAI announce a 10-gigawatt letter of intent", "summary": "Nvidia said it intends to invest up to $100 billion in OpenAI as each gigawatt of its systems is deployed.", "primary_source": "https://nvidianews.nvidia.com/news/openai-and-nvidia-announce-strategic-partnership-to-deploy-10gw-of-nvidia-systems", "secondary_source": "https://fortune.com/2025/12/02/nvidia-openai-deal-not-signed-yet-100-billion-rally-colette-kress" }, { "date": "2025-09-22", "date_label": "", "lab": "nvidia", "type": "infrastructure", "title": "OpenAI and Nvidia 10 GW partnership", "summary": "OpenAI and Nvidia announced a letter of intent to deploy at least 10 gigawatts of Nvidia systems.", "primary_source": "https://nvidianews.nvidia.com/news/openai-and-nvidia-announce-strategic-partnership-to-deploy-10gw-of-nvidia-systems", "secondary_source": "" }, { "date": "2025-09-23", "date_label": "", "lab": "openai", "type": "infrastructure", "title": "Stargate adds five sites", "summary": "OpenAI, Oracle and SoftBank announced five new Stargate sites, bringing planned capacity to nearly 7 gigawatts.", "primary_source": "https://openai.com/index/five-new-stargate-sites/", "secondary_source": "" }, { "date": "2025-09-26", "date_label": "", "lab": "telus", "type": "infrastructure", "title": "Telus opens its Rimouski AI factory", "summary": "Telus opened a sovereign AI factory in Rimouski, Quebec, using Nvidia GPUs and HPE infrastructure.", "primary_source": "https://rcrwireless.com/20250926/ai-infrastructure/telus-ai-factory", "secondary_source": "https://www.rcrwireless.com/20260512/carriers/telus-ai-data-centers" }, { "date": "2025-09-29", "date_label": "", "lab": "anthropic", "type": "model", "title": "Claude Sonnet 4.5", "summary": "Anthropic released Sonnet 4.5, focused on coding and long-running agents, with the Claude Agent SDK.", "primary_source": "https://www.anthropic.com/news/claude-sonnet-4-5", "secondary_source": "https://www.axios.com/2025/09/29/anthropic-claude-sonnet-coding-agent" }, { "date": "2025-09-29", "date_label": "", "lab": "deepseek", "type": "model", "title": "DeepSeek-V3.2-Exp", "summary": "DeepSeek released the experimental V3.2.", "primary_source": "https://api-docs.deepseek.com/news/news260910", "secondary_source": "https://huggingface.co/deepseek-ai/DeepSeek-V3.2-Exp" }, { "date": "2025-09-29", "date_label": "", "lab": "us-commerce", "type": "policy", "title": "US extends Entity List to affiliates", "summary": "The US Commerce Department extended end-user controls to affiliates of listed entities.", "primary_source": "https://www.federalregister.gov/documents/2025/09/30/2025-19001/expansion-of-end-user-controls-to-cover-affiliates-of-certain-listed-entities", "secondary_source": "" }, { "date": "2025-09-30", "date_label": "", "lab": "openai", "type": "model", "title": "Sora 2 and the Sora app", "summary": "OpenAI released Sora 2, a video model with synchronized audio, alongside an invite-only social iOS app with cameos.", "primary_source": "https://openai.com/index/sora-2/", "secondary_source": "https://gigazine.net/gsc_news/en/20251001-openai-sora-for-ios" }, { "date": "2025-09-30", "date_label": "", "lab": "netherlands", "type": "policy", "title": "Netherlands invokes emergency powers over Nexperia", "summary": "The Dutch government invoked the Goods Availability Act over the Chinese-owned chipmaker Nexperia, over fears of technology transfer.", "primary_source": "https://www.euronews.com/business/2025/11/19/dutch-government-suspends-takeover-of-chinese-owned-nexperia", "secondary_source": "https://www.tomshardware.com/tech-industry/semiconductors/netherlands-suspends-nexperia-takeover-order-as-china-eases-export-curbs" }, { "date": "2025-10-01", "date_label": "", "lab": "tsmc", "type": "chip", "title": "TSMC starts 2 nm volume production", "summary": "TSMC says its 2-nanometre process, its first with nanosheet transistors, started volume production in the fourth quarter of 2025.", "primary_source": "https://www.tsmc.com/english/dedicatedFoundry/technology/logic/l_2nm", "secondary_source": "" }, { "date": "2025-10-06", "date_label": "", "lab": "openai", "type": "api", "title": "AgentKit and Agent Builder", "summary": "OpenAI released AgentKit, including the visual Agent Builder, a connector registry and ChatKit for embedding agents.", "primary_source": "https://openai.com/index/introducing-agentkit/", "secondary_source": "https://openai.com/devday/2025/" }, { "date": "2025-10-06", "date_label": "", "lab": "openai", "type": "api", "title": "Apps in ChatGPT and the Apps SDK", "summary": "OpenAI let third-party apps run inside ChatGPT conversations, built on the Model Context Protocol.", "primary_source": "https://openai.com/index/introducing-apps-in-chatgpt/", "secondary_source": "https://en.wikipedia.org/wiki/ChatGPT" }, { "date": "2025-10-06", "date_label": "", "lab": "openai", "type": "event", "title": "OpenAI DevDay 2025", "summary": "OpenAI announced apps inside ChatGPT with the Apps SDK, AgentKit, general availability of Codex, and Sora 2 and GPT-5 Pro in the API.", "primary_source": "https://openai.com/devday/2025/", "secondary_source": "https://community.openai.com/t/devday-2025-apps-sdk-sora-2-gpt-5-pro-agentkit-new-image-generation-and-speech-to-speech-mini-models-and-more/1361279" }, { "date": "2025-10-06", "date_label": "", "lab": "amd", "type": "chip", "title": "AMD and OpenAI sign a 6 GW GPU agreement", "summary": "OpenAI agreed to deploy up to 6 GW of AMD GPUs and received a warrant for up to 160 million AMD shares.", "primary_source": "https://ir.amd.com/news-events/press-releases/detail/1260/amd-and-openai-announce-strategic-partnership-to-deploy-6-gigawatts-of-amd-gpus", "secondary_source": "https://www.sec.gov/Archives/edgar/data/0000002488/000000248826000123/amd-20260627.htm" }, { "date": "2025-10-06", "date_label": "", "lab": "openai", "type": "finance", "title": "AMD gives OpenAI a warrant in a 6-gigawatt chip deal", "summary": "AMD issued OpenAI a warrant for up to 160 million shares that vests as OpenAI deploys up to 6 gigawatts of AMD GPUs.", "primary_source": "https://ir.amd.com/news-events/press-releases/detail/1260/amd-and-openai-announce-strategic-partnership-to-deploy-6-gigawatts-of-amd-gpus", "secondary_source": "" }, { "date": "2025-10-09", "date_label": "", "lab": "china", "type": "policy", "title": "China widens rare earth export controls", "summary": "China issued six announcements widening export controls, including on rare earths.", "primary_source": "https://www.pillsburylaw.com/en/news-and-insights/china-suspends-export-controls-certain-critical-minerals-related-items.html", "secondary_source": "https://www.china-briefing.com/news/chinas-rare-earth-export-controls-impacts-on-businesses/" }, { "date": "2025-10-09", "date_label": "", "lab": "microsoft", "type": "infrastructure", "title": "Microsoft's first large GB300 cluster", "summary": "Microsoft Azure said it had delivered the first at-scale production cluster with more than 4,600 GB300 NVL72 systems, for OpenAI workloads.", "primary_source": "https://azure.microsoft.com/en-us/blog/microsoft-azure-delivers-the-first-large-scale-cluster-with-nvidia-gb300-nvl72-for-openai-workloads/", "secondary_source": "" }, { "date": "2025-10-13", "date_label": "", "lab": "nvidia", "type": "hardware", "title": "Nvidia DGX Spark ships", "summary": "Nvidia said it and its partners were shipping DGX Spark that week, with 128 GB of unified memory and a claimed petaflop of AI performance, and that orders would open on October 15.", "primary_source": "https://nvidianews.nvidia.com/news/nvidia-dgx-spark-arrives-for-worlds-ai-developers", "secondary_source": "" }, { "date": "2025-10-13", "date_label": "", "lab": "openai", "type": "chip", "title": "OpenAI and Broadcom custom accelerators", "summary": "OpenAI and Broadcom announced a collaboration to deploy 10 gigawatts of OpenAI-designed accelerators.", "primary_source": "https://openai.com/index/openai-and-broadcom-announce-strategic-collaboration/", "secondary_source": "" }, { "date": "2025-10-14", "date_label": "", "lab": "intel", "type": "chip", "title": "Intel announces the Crescent Island inference GPU", "summary": "Intel announced a data-centre GPU for inference, using Xe3P and 160 GB of LPDDR5X, with customer samples expected in the second half of 2026.", "primary_source": "https://www.intel.com/content/www/us/en/newsroom/news/artificial-intelligence/intel-to-expand-ai-accelerator-portfolio-with-new-gpu.html", "secondary_source": "https://www.datacenterdynamics.com/en/news/intel-unveils-crescent-island-data-center-gpu-for-inferencing-workloads/" }, { "date": "2025-10-15", "date_label": "", "lab": "anthropic", "type": "model", "title": "Claude Haiku 4.5", "summary": "Anthropic released Haiku 4.5, a small fast model.", "primary_source": "https://www.anthropic.com/news/claude-haiku-4-5", "secondary_source": "https://en.wikipedia.org/wiki/Claude_(language_model)" }, { "date": "2025-10-15", "date_label": "", "lab": "apple", "type": "chip", "title": "Apple M5", "summary": "Apple announced the M5, with 153 GB per second of unified memory bandwidth.", "primary_source": "https://www.apple.com/newsroom/2025/10/apple-unleashes-m5-the-next-big-leap-in-ai-performance-for-apple-silicon/", "secondary_source": "" }, { "date": "2025-10-21", "date_label": "", "lab": "openai", "type": "product", "title": "ChatGPT Atlas browser", "summary": "OpenAI launched ChatGPT Atlas, a macOS web browser with ChatGPT built in and an agent mode for paid users.", "primary_source": "https://openai.com/index/introducing-chatgpt-atlas/", "secondary_source": "https://gulfnews.com/technology/media/openai-launches-chatgptatlas-the-ai-powered-browser-that-thinks-with-you-1.500316403" }, { "date": "2025-10-21", "date_label": "", "lab": "meta", "type": "finance", "title": "Meta and Blue Owl form a $27 billion Hyperion venture", "summary": "Meta kept 20% of a joint venture that will own its Louisiana campus, with Blue Owl funds holding 80% and costs of about $27 billion.", "primary_source": "https://about.fb.com/news/2025/10/meta-blue-owl-capital-develop-hyperion-data-center/", "secondary_source": "https://www.sec.gov/Archives/edgar/data/1326801/000162828026066773/meta-20260930.htm" }, { "date": "2025-10-23", "date_label": "", "lab": "anthropic", "type": "infrastructure", "title": "Anthropic plans up to one million Google TPUs", "summary": "Anthropic said it would expand its use of Google Cloud with up to one million TPUs, in a deal worth tens of billions of dollars.", "primary_source": "https://www.anthropic.com/news/expanding-our-use-of-google-cloud-tpus-and-services", "secondary_source": "" }, { "date": "2025-10-27", "date_label": "", "lab": "qualcomm", "type": "chip", "title": "Qualcomm AI200 and AI250", "summary": "Qualcomm announced AI200 and AI250 inference accelerators and racks for data centres, due in 2026 and 2027.", "primary_source": "https://www.edge-ai-vision.com/2025/10/qualcomm-unveils-ai200-and-ai250-redefining-rack-scale-data-center-inference-performance-for-the-ai-era/", "secondary_source": "" }, { "date": "2025-10-28", "date_label": "", "lab": "us-doe", "type": "hardware", "title": "US announces Solstice and Equinox at Argonne", "summary": "The DOE, Nvidia and Oracle announced Solstice, with 100,000 Blackwell GPUs, and Equinox, with 10,000, at Argonne.", "primary_source": "https://www.energy.gov/articles/energy-department-announces-new-partnership-nvidia-and-oracle-build-largest-doe-ai", "secondary_source": "https://www.theregister.com/2025/10/28/nvidia_oracle_supercomputers_doe/" }, { "date": "2025-11", "date_label": "", "lab": "canada", "type": "infrastructure", "title": "Alberta allocates its 1,200 MW data center limit", "summary": "Alberta's grid operator allocated its interim 1,200 MW limit to two projects, leaving 37 proposals seeking 19.4 GW.", "primary_source": "https://www.cbc.ca/news/canada/calgary/ai-data-centre-alberta-electricity-9.6977136", "secondary_source": "https://www.blg.com/en/insights/2026/07/data-centre-regulation-in-british-columbia-competing-for-a-limited-supply" }, { "date": "2025-11-04", "date_label": "", "lab": "google", "type": "infrastructure", "title": "Google unveils Project Suncatcher", "summary": "Google Research published a design for solar-powered satellite constellations carrying TPUs, with two prototypes planned with Planet by early 2027.", "primary_source": "https://research.google/blog/exploring-a-space-based-scalable-ai-infrastructure-system-design/", "secondary_source": "https://services.google.com/fh/files/misc/suncatcher_paper.pdf" }, { "date": "2025-11-06", "date_label": "", "lab": "google", "type": "product", "title": "Google says its Ironwood TPU will be generally available within weeks", "summary": "Google announced that Ironwood, its seventh-generation TPU, would be generally available in the coming weeks, alongside new Axion virtual machines.", "primary_source": "https://cloud.google.com/blog/products/compute/ironwood-tpus-and-new-axion-based-vms-for-your-ai-workloads", "secondary_source": "https://www.constellationr.com/insights/news/google-clouds-ironwood-ready-general-availability" }, { "date": "2025-11-07", "date_label": "", "lab": "china", "type": "policy", "title": "China pauses its October export controls", "summary": "China's commerce ministry suspended the six October 9 announcements for one year.", "primary_source": "https://www.pillsburylaw.com/en/news-and-insights/china-suspends-export-controls-certain-critical-minerals-related-items.html", "secondary_source": "https://www.fdd.org/analysis/2025/11/12/china-pauses-some-rare-earth-export-curbs-while-retaining-levers-of-control/" }, { "date": "2025-11-10", "date_label": "", "lab": "us-commerce", "type": "policy", "title": "US suspends the affiliates rule for a year", "summary": "The US suspended its rule extending export limits to affiliates of listed entities, until November 9, 2026.", "primary_source": "https://www.federalregister.gov/api/v1/documents/2025-19846.json", "secondary_source": "https://www.wsgr.com/en/insights/bis-formally-suspends-new-affiliates-rule-for-one-year.html" }, { "date": "2025-11-12", "date_label": "", "lab": "openai", "type": "model", "title": "GPT-5.1 Instant and Thinking", "summary": "OpenAI updated ChatGPT to GPT-5.1, aiming for warmer conversation, adaptive reasoning and new personality presets.", "primary_source": "https://openai.com/index/gpt-5-1/", "secondary_source": "https://arstechnica.com/ai/2025/11/openai-walks-a-tricky-tightrope-with-gpt-5-1s-eight-new-personalities/" }, { "date": "2025-11-13", "date_label": "", "lab": "openai", "type": "feature", "title": "Group chats pilot", "summary": "OpenAI piloted group chats where up to 20 people can share one ChatGPT conversation, starting in a few countries.", "primary_source": "https://openai.com/index/group-chats-in-chatgpt/", "secondary_source": "https://help.openai.com/en/articles/6825453-chatgpt-release-notes" }, { "date": "2025-11-18", "date_label": "", "lab": "google", "type": "model", "title": "Gemini 3 Pro", "summary": "Google released Gemini 3 Pro in the Gemini app, Search and developer tools.", "primary_source": "https://storage.googleapis.com/deepmind-media/Model-Cards/Gemini-3-Pro-Model-Card.pdf", "secondary_source": "https://en.wikipedia.org/wiki/Gemini_(language_model)" }, { "date": "2025-11-18", "date_label": "", "lab": "anthropic", "type": "finance", "title": "Microsoft and Nvidia commit to invest in Anthropic", "summary": "Anthropic committed to $30 billion of Azure compute while Nvidia and Microsoft committed up to $10 billion and $5 billion of investment.", "primary_source": "https://blogs.microsoft.com/blog/2025/11/18/microsoft-nvidia-and-anthropic-announce-strategic-partnerships/", "secondary_source": "" }, { "date": "2025-11-18", "date_label": "", "lab": "eurohpc", "type": "hardware", "title": "Contract signed for Alice Recoque", "summary": "EuroHPC and Eviden signed the contract for Alice Recoque, Europe's second exascale system, to be hosted in France.", "primary_source": "https://digital-strategy.ec.europa.eu/en/news/alice-recoque-europes-second-exascale-supercomputer-its-way", "secondary_source": "https://hwbusters.com/news/amd-and-eviden-build-frances-first-exascale-supercomputer/" }, { "date": "2025-11-18", "date_label": "", "lab": "anthropic", "type": "infrastructure", "title": "Anthropic, Microsoft and Nvidia partnership", "summary": "Anthropic committed to buy 30 billion US dollars of Azure compute and up to one gigawatt of additional capacity, in a partnership with Microsoft and Nvidia.", "primary_source": "https://blogs.microsoft.com/blog/2025/11/18/microsoft-nvidia-and-anthropic-announce-strategic-partnerships/", "secondary_source": "" }, { "date": "2025-11-19", "date_label": "", "lab": "openai", "type": "model", "title": "GPT-5.1-Codex-Max", "summary": "OpenAI released a coding model for long-running work that compacts its own context to keep going across sessions.", "primary_source": "https://openai.com/index/gpt-5-1-codex-max/", "secondary_source": "https://venturebeat.com/ai/openai-debuts-gpt-5-1-codex-max-coding-model-and-it-already-completed-a-24" }, { "date": "2025-11-20", "date_label": "", "lab": "google", "type": "model", "title": "Nano Banana Pro (Gemini 3 Pro Image)", "summary": "Google released an image model built on Gemini 3 Pro with better text rendering.", "primary_source": "https://blog.google/technology/ai/nano-banana-pro/", "secondary_source": "https://en.wikipedia.org/wiki/Gemini_(language_model)" }, { "date": "2025-11-24", "date_label": "", "lab": "openai", "type": "feature", "title": "Shopping research in ChatGPT", "summary": "ChatGPT gained a shopping research mode that builds buyer's guides using a model based on GPT-5 thinking mini.", "primary_source": "https://openai.com/index/chatgpt-shopping-research/", "secondary_source": "https://en.wikipedia.org/wiki/GPT-5" }, { "date": "2025-11-24", "date_label": "", "lab": "anthropic", "type": "model", "title": "Claude Opus 4.5", "summary": "Anthropic released Opus 4.5 with gains in coding and office work at lower prices.", "primary_source": "https://www.anthropic.com/news/claude-opus-4-5", "secondary_source": "https://www.tomsguide.com/ai/claude-opus-4-5-is-here-anthropics-most-powerful-version-to-date" }, { "date": "2025-11-24", "date_label": "", "lab": "us-gov", "type": "policy", "title": "Trump signs the Genesis Mission order", "summary": "An executive order launched the Genesis Mission, a Department of Energy effort to apply national computing and data to AI for science.", "primary_source": "https://www.whitehouse.gov/presidential-actions/2025/11/launching-the-genesis-mission/", "secondary_source": "https://ktslaw.com/en/Insights/Alert/2025/11/Genesis-Mission-Artificial-Intelligence-Executive-Order" }, { "date": "2025-12-01", "date_label": "", "lab": "deepseek", "type": "model", "title": "DeepSeek-V3.2", "summary": "DeepSeek released V3.2 with sparse attention for cheaper long-context use.", "primary_source": "https://api-docs.deepseek.com/news/news251201", "secondary_source": "https://en.wikipedia.org/wiki/DeepSeek" }, { "date": "2025-12-02", "date_label": "", "lab": "mistral", "type": "model", "title": "Mistral 3 and Mistral Large 3", "summary": "Mistral released the Mistral 3 family, including Mistral Large 3 and small Ministral 3 models, under open licences.", "primary_source": "https://mistral.ai/news/mistral-3", "secondary_source": "https://en.wikipedia.org/wiki/Mistral_AI" }, { "date": "2025-12-02", "date_label": "", "lab": "amazon", "type": "chip", "title": "AWS Trainium3 and Trn3 UltraServers", "summary": "AWS launched Trainium3, its first 3-nanometre AI chip, in UltraServers of up to 144 chips, and said it was already working on Trainium4.", "primary_source": "https://www.aboutamazon.com/news/aws/trainium-3-ultraserver-faster-ai-training-lower-cost", "secondary_source": "" }, { "date": "2025-12-03", "date_label": "", "lab": "micron", "type": "hardware", "title": "Micron announces exit from its Crucial consumer business", "summary": "Micron said it would leave the Crucial consumer brand to better supply larger strategic customers amid AI demand.", "primary_source": "https://investors.micron.com/news-releases/news-release-details/micron-announces-exit-crucial-consumer-business", "secondary_source": "" }, { "date": "2025-12-11", "date_label": "", "lab": "openai", "type": "model", "title": "GPT-5.2", "summary": "OpenAI released GPT-5.2 in Instant, Thinking and Pro versions, focused on professional work and long tasks.", "primary_source": "https://openai.com/index/introducing-gpt-5-2/", "secondary_source": "https://www.theverge.com/ai-artificial-intelligence/842529/openai-gpt-5-2-new-model-chatgpt" }, { "date": "2025-12-16", "date_label": "", "lab": "openai", "type": "model", "title": "New ChatGPT Images (GPT Image 1.5)", "summary": "OpenAI rolled out a new image model to all ChatGPT users and the API, with more precise edits and faster generation.", "primary_source": "https://openai.com/index/new-chatgpt-images-is-here/", "secondary_source": "https://community.openai.com/t/gpt-image-1-5-rolling-out-in-the-api-and-chatgpt/1369443" }, { "date": "2025-12-17", "date_label": "", "lab": "google", "type": "model", "title": "Gemini 3 Flash", "summary": "Google released Gemini 3 Flash, a faster lower-cost Gemini 3 model.", "primary_source": "https://blog.google/products/gemini/gemini-3-flash/", "secondary_source": "https://en.wikipedia.org/wiki/Gemini_(language_model)" }, { "date": "2025-12-18", "date_label": "", "lab": "openai", "type": "model", "title": "GPT-5.2-Codex", "summary": "OpenAI released a GPT-5.2 version optimized for agentic coding in Codex, with API access added in January 2026.", "primary_source": "https://openai.com/index/introducing-gpt-5-2-codex/", "secondary_source": "https://venturebeat.com/technology/enterprise-ai-coding-grows-teeth-gpt-5-2-codex-weaves-security-into-large" }, { "date": "2025-12-24", "date_label": "", "lab": "groq", "type": "chip", "title": "Groq licenses inference technology to Nvidia", "summary": "Groq announced a non-exclusive licensing agreement giving Nvidia access to its inference technology.", "primary_source": "https://groq.com/newsroom/groq-and-nvidia-enter-non-exclusive-inference-technology-licensing-agreement-to-accelerate-ai-inference-at-global-scale", "secondary_source": "" }, { "date": "2026-01", "date_label": "", "lab": "spacex", "type": "infrastructure", "title": "SpaceX applies for up to one million orbital data center satellites", "summary": "SpaceX applied to the US FCC for up to one million satellites intended as a distributed data center.", "primary_source": "https://www.cnbc.com/2026/06/21/do-space-based-ai-data-centers-make-economic-sense.html", "secondary_source": "https://www.techcrunch.com/2026/03/20/jeff-bezos-blue-origin-enters-the-space-data-center-game/" }, { "date": "2026-01", "date_label": "", "lab": "moonshot", "type": "model", "title": "Kimi K2.5", "summary": "Moonshot added native vision and agent swarms of up to 100 sub-agents to the K2 line in Kimi K2.5.", "primary_source": "https://www.kimi.com/blog/kimi-k2-5", "secondary_source": "https://huggingface.co/moonshotai/Kimi-K2.5" }, { "date": "2026-01-05", "date_label": "", "lab": "nvidia", "type": "chip", "title": "Nvidia details the Rubin platform", "summary": "At CES, Nvidia detailed Rubin as six new chips forming one AI supercomputer, with partner products due in the second half of 2026.", "primary_source": "https://nvidianews.nvidia.com/news/rubin-platform-ai-supercomputer", "secondary_source": "" }, { "date": "2026-01-05", "date_label": "", "lab": "amd", "type": "hardware", "title": "AMD previews Helios at CES", "summary": "At CES, AMD showed its Helios rack-scale platform and the Instinct MI455X accelerator.", "primary_source": "https://newsroom.amd.com/news/amd-and-its-partners-share-their-vision-for-ai-ev/", "secondary_source": "" }, { "date": "2026-01-12", "date_label": "", "lab": "apple", "type": "event", "title": "Apple and Google Gemini partnership", "summary": "Apple and Google said the next Apple Foundation Models would be built on Gemini models and cloud technology.", "primary_source": "https://blog.google/company-news/inside-google/company-announcements/joint-statement-google-apple/", "secondary_source": "https://www.mactech.com/2026/01/12/apple-will-use-googles-gemini-for-its-ai-powered-siri" }, { "date": "2026-01-12", "date_label": "", "lab": "anthropic", "type": "product", "title": "Claude Cowork research preview", "summary": "Anthropic previewed Cowork, an agent in the Claude desktop app that works on local files for non-programmers, starting with Max subscribers.", "primary_source": "https://claude.com/blog/cowork-research-preview", "secondary_source": "https://www.wired.com/story/anthropic-claude-cowork-agent/" }, { "date": "2026-01-13", "date_label": "", "lab": "us-bis", "type": "policy", "title": "US moves H200-class chips to case-by-case review for China", "summary": "BIS changed licence review for chips such as the Nvidia H200 and AMD MI325X from presumption of denial to case-by-case, under conditions.", "primary_source": "https://www.bis.gov/press-release/department-commerce-revises-license-review-policy-semiconductors-exported-china", "secondary_source": "https://public-inspection.federalregister.gov/2026-00789.pdf" }, { "date": "2026-01-14", "date_label": "", "lab": "us-gov", "type": "policy", "title": "Trump sets a 25% tariff on certain advanced chips", "summary": "A proclamation placed a 25% tariff on a narrow group of advanced chips, with exemptions such as use in US data centres.", "primary_source": "https://www.whitehouse.gov/presidential-actions/2026/01/adjusting-imports-of-semiconductors-semiconductor-manufacturing-equipment-and-their-derivative-products-into-the-united-states/", "secondary_source": "https://www.thomsonreuters.com/en-us/help/onesource-global-trade/regulatory-insights/2026/january-15th/trump-admin-tariffs-advanced-ai-chips-232.html" }, { "date": "2026-01-14", "date_label": "", "lab": "openai", "type": "infrastructure", "title": "OpenAI adds 750 MW of Cerebras compute", "summary": "OpenAI said it would add 750 megawatts of low-latency Cerebras compute to its platform.", "primary_source": "https://openai.com/index/cerebras-partnership/", "secondary_source": "" }, { "date": "2026-01-15", "date_label": "", "lab": "us-commerce", "type": "policy", "title": "US rule moves H200-class chips to case-by-case review", "summary": "A Commerce rule moved exports of Nvidia H200-class chips to China and Macau from a presumption of denial to case-by-case review.", "primary_source": "https://www.federalregister.gov/documents/2026/01/15/2026-00789/revision-to-license-review-policy-for-advanced-computing-commodities", "secondary_source": "https://www.mayerbrown.com/en/insights/publications/2026/01/administration-policies-on-advanced-ai-chips-codified" }, { "date": "2026-01-16", "date_label": "", "lab": "openai", "type": "product", "title": "ChatGPT Go worldwide; ad tests announced", "summary": "OpenAI made the $8 ChatGPT Go plan available everywhere and said it would test labeled ads for US Free and Go users.", "primary_source": "https://openai.com/index/introducing-chatgpt-go/", "secondary_source": "https://www.thurrott.com/cloud/331745/chatgpt-go-is-now-available-worldwide-but-free-and-go-users-will-soon-see-ads" }, { "date": "2026-01-16", "date_label": "", "lab": "eu", "type": "policy", "title": "EU Council clears AI gigafactories", "summary": "The Council extended EuroHPC's mandate to cover AI gigafactories.", "primary_source": "https://transition-pathways.europa.eu/adv-man/news/council-clears-way-ai-gigafactories-across-eu", "secondary_source": "https://openfuture.eu/blog/the-shaky-sovereignty-bet-behind-the-eus-ai-gigafactories/" }, { "date": "2026-01-26", "date_label": "", "lab": "microsoft", "type": "chip", "title": "Microsoft Maia 200", "summary": "Microsoft announced Maia 200, an inference accelerator on TSMC's 3-nanometre process with 216 GB of HBM3e memory.", "primary_source": "https://blogs.microsoft.com/blog/2026/01/26/maia-200-the-ai-accelerator-built-for-inference/", "secondary_source": "" }, { "date": "2026-01-29", "date_label": "", "lab": "openai", "type": "retirement", "title": "Retirement of GPT-4o in ChatGPT announced", "summary": "OpenAI said GPT-4o, GPT-4.1, GPT-4.1 mini and o4-mini would leave ChatGPT on 13 February 2026, with the API unchanged.", "primary_source": "https://openai.com/index/retiring-gpt-4o-and-older-models/", "secondary_source": "https://yourstory.com/ai-story/openai-retiring-gpt-4o-older-models-chatgpt" }, { "date": "2026-01-30", "date_label": "", "lab": "canada", "type": "policy", "title": "BC launches competitive call for data center power", "summary": "BC Hydro and the province launched a call under limits of 100 MW for conventional and 300 MW for AI data centers over two years.", "primary_source": "https://www.blg.com/en/insights/2026/07/data-centre-regulation-in-british-columbia-competing-for-a-limited-supply", "secondary_source": "https://gowlingwlg.com/en/insights-resources/articles/2026/power-shifts-canada-data-centre-market" }, { "date": "2026-02", "date_label": "", "lab": "india", "type": "event", "title": "India adds GPUs at the AI Impact Summit", "summary": "At the AI Impact Summit, India announced 20,000 more GPUs on top of 38,000 provisioned.", "primary_source": "https://www.pib.gov.in/PressReleasePage.aspx?PRID=2234343®=3&lang=1", "secondary_source": "" }, { "date": "2026-02-01", "date_label": "", "lab": "us-commerce", "type": "policy", "title": "US grants Nvidia a first H200 licence for China", "summary": "Nvidia's annual report says the US government granted a February 2026 licence to ship small amounts of H200 chips to specific China-based customers, with US inspection and a 25% tariff.", "primary_source": "https://www.sec.gov/Archives/edgar/data/1045810/000104581026000021/nvda-20260125.htm", "secondary_source": "https://bloomberg.com/news/articles/2026-02-26/nvidia-gets-us-license-for-small-amount-of-h200-exports-to-china" }, { "date": "2026-02-02", "date_label": "", "lab": "openai", "type": "product", "title": "Codex app for macOS", "summary": "OpenAI released a desktop Codex app for managing several coding agents in parallel, with automations; Windows followed on 4 March.", "primary_source": "https://openai.com/index/introducing-the-codex-app/", "secondary_source": "https://siliconangle.com/2026/02/02/openai-launches-codex-app-macos-speed-software-projects/" }, { "date": "2026-02-02", "date_label": "", "lab": "xai", "type": "finance", "title": "xAI becomes part of SpaceX", "summary": "SpaceX absorbed xAI, making it a wholly owned subsidiary, according to its later IPO filing.", "primary_source": "https://www.sec.gov/Archives/edgar/data/0001181412/000162828026036936/spaceexplorationtechnologi.htm", "secondary_source": "" }, { "date": "2026-02-05", "date_label": "", "lab": "openai", "type": "model", "title": "GPT-5.3-Codex", "summary": "OpenAI released GPT-5.3-Codex, a faster coding model pitched for broader computer work beyond code.", "primary_source": "https://openai.com/index/introducing-gpt-5-3-codex/", "secondary_source": "https://www.zdnet.com/article/openai-gpt-5-3-codex-faster-goes-beyond-coding/" }, { "date": "2026-02-05", "date_label": "", "lab": "anthropic", "type": "model", "title": "Claude Opus 4.6", "summary": "Anthropic released Opus 4.6 with agent teams and Claude in PowerPoint.", "primary_source": "https://www.anthropic.com/news/claude-opus-4-6", "secondary_source": "https://techcrunch.com/2026/02/05/anthropic-releases-opus-4-6-with-new-agent-teams/" }, { "date": "2026-02-12", "date_label": "", "lab": "anthropic", "type": "finance", "title": "Anthropic raises $30 billion at $380 billion", "summary": "Anthropic announced a $30 billion Series G led by GIC and Coatue at a $380 billion post-money valuation.", "primary_source": "https://www.anthropic.com/news/anthropic-raises-30-billion-series-g-funding-380-billion-post-money-valuation", "secondary_source": "https://www.sec.gov/Archives/edgar/data/2130357/000162828026067168/anthropicpbc-sx1.htm" }, { "date": "2026-02-12", "date_label": "", "lab": "zai", "type": "model", "title": "GLM-5", "summary": "Z.ai released GLM-5 under MIT, growing to 744 billion parameters with 40 billion active.", "primary_source": "https://huggingface.co/zai-org/GLM-5", "secondary_source": "https://en.wikipedia.org/wiki/Z.ai" }, { "date": "2026-02-12", "date_label": "", "lab": "samsung", "type": "chip", "title": "Samsung ships commercial HBM4", "summary": "Samsung said it had started shipping commercial HBM4, at 11.7 gigabits per second.", "primary_source": "https://news.samsung.com/global/samsung-ships-industry-first-commercial-hbm4-with-ultimate-performance-for-ai-computing", "secondary_source": "" }, { "date": "2026-02-13", "date_label": "", "lab": "openai", "type": "retirement", "title": "GPT-4o retired from ChatGPT", "summary": "GPT-4o and several older models were removed from ChatGPT; business plans kept GPT-4o in custom GPTs until 3 April 2026.", "primary_source": "https://help.openai.com/en/articles/20001051-retiring-gpt-4o-and-other-chatgpt-models", "secondary_source": "https://letsdatascience.com/news/openai-retires-gpt-4o-from-chatgpt-5a9102f8" }, { "date": "2026-02-16", "date_label": "", "lab": "alibaba", "type": "model", "title": "Qwen3.5", "summary": "Alibaba released Qwen3.5.", "primary_source": "https://qwen.ai/blog?id=qwen3.5", "secondary_source": "https://en.wikipedia.org/wiki/Qwen" }, { "date": "2026-02-17", "date_label": "", "lab": "anthropic", "type": "model", "title": "Claude Sonnet 4.6", "summary": "Anthropic released Sonnet 4.6, a faster and cheaper model.", "primary_source": "https://www.anthropic.com/news/claude-sonnet-4-6", "secondary_source": "https://www.axios.com/2026/02/17/anthropic-new-claude-sonnet-faster-cheaper" }, { "date": "2026-02-19", "date_label": "", "lab": "google", "type": "model", "title": "Gemini 3.1 Pro", "summary": "Google released Gemini 3.1 Pro.", "primary_source": "https://deepmind.google/models/model-cards/gemini-3-1-pro/", "secondary_source": "https://en.wikipedia.org/wiki/Gemini_(language_model)" }, { "date": "2026-02-19", "date_label": "", "lab": "canada", "type": "policy", "title": "Hydro-Québec proposes a 13-cent data center rate", "summary": "Hydro-Québec asked the Régie de l'énergie to approve an average rate of 13 cents/kWh for data centers above 5 MW.", "primary_source": "https://news.hydroquebec.com/news/press-releases/all-quebec/hydro-quebec-proposing-regie-energie-new-rate-large-data-centres-adjustment-rate-cryptographic-use-applied-blockchains.html", "secondary_source": "https://www.lenouvelliste.ca/actualites/2026/10/01/centres-de-donnees-debut-dune-audience-sur-les-prix-de-lelectricite-4XS4KX52GFG2FETIEOOECFIQSI/" }, { "date": "2026-03", "date_label": "", "lab": "openai", "type": "infrastructure", "title": "OpenAI and Oracle drop the Abilene expansion", "summary": "OpenAI reversed its plan to expand Abilene to 2.1 GW, and Microsoft partnered with Crusoe on the adjacent 900 MW.", "primary_source": "https://epochai.substack.com/p/openai-stargate-where-the-us-sites", "secondary_source": "https://www.sherwood.news/tech/openais-stargate-shrinks-further-as-uk-data-center-paused/" }, { "date": "2026-03", "date_label": "", "lab": "aws", "type": "event", "title": "Drone strikes hit AWS data centres in the Gulf", "summary": "Drones struck two Amazon data centres in the UAE and a nearby strike damaged one in Bahrain during the Iran war.", "primary_source": "https://www.cnbc.com/2026/09/15/aws-cant-restore-service-to-bahrain-uae-6-months-after-iran-strikes.html", "secondary_source": "https://asiapolicy.asiasociety.org/p/iran-is-hitting-data-centers-in-the" }, { "date": "2026-03-03", "date_label": "", "lab": "openai", "type": "model", "title": "GPT-5.3 Instant", "summary": "OpenAI made GPT-5.3 Instant the ChatGPT default, aiming for fewer needless refusals and better web answers.", "primary_source": "https://openai.com/index/gpt-5-3-instant-system-card/", "secondary_source": "https://gigazine.net/gsc_news/en/20260304-gpt-5-3-instant" }, { "date": "2026-03-03", "date_label": "", "lab": "apple", "type": "chip", "title": "Apple M5 Pro and M5 Max", "summary": "Apple debuted the M5 Pro and M5 Max, with unified memory bandwidth up to 614 GB per second.", "primary_source": "https://www.apple.com/newsroom/2026/03/apple-debuts-m5-pro-and-m5-max-to-supercharge-the-most-demanding-pro-workflows/", "secondary_source": "" }, { "date": "2026-03-05", "date_label": "", "lab": "openai", "type": "model", "title": "GPT-5.4 Thinking and Pro", "summary": "OpenAI released GPT-5.4 in Thinking and Pro versions for paid ChatGPT users and the API.", "primary_source": "https://openai.com/index/introducing-gpt-5-4/", "secondary_source": "https://techcrunch.com/2026/03/05/openai-launches-gpt-5-4-with-pro-and-thinking-versions/" }, { "date": "2026-03-11", "date_label": "", "lab": "meta", "type": "chip", "title": "Meta four-chip MTIA roadmap", "summary": "Meta said it was developing four MTIA generations, with MTIA 300 already in production for ranking and recommendation training.", "primary_source": "https://about.fb.com/news/2026/03/expanding-metas-custom-silicon-to-power-our-AI-workloads/", "secondary_source": "" }, { "date": "2026-03-16", "date_label": "", "lab": "micron", "type": "chip", "title": "Micron in high-volume HBM4 production", "summary": "Micron said its 36 GB 12-high HBM4, designed for Nvidia's Vera Rubin, was in high-volume production at over 2.8 TB/s.", "primary_source": "https://investors.micron.com/news-releases/news-release-details/micron-high-volume-production-hbm4-designed-nvidia-vera-rubin", "secondary_source": "https://www.sec.gov/Archives/edgar/data/0000723125/000072312526000023/mu-20260903.htm" }, { "date": "2026-03-16", "date_label": "", "lab": "nvidia", "type": "chip", "title": "Vera Rubin at GTC 2026, with the Groq 3 LPX rack", "summary": "At GTC, Nvidia presented the Vera Rubin platform, including the Vera Rubin NVL72 GPU rack and the Groq 3 LPX inference accelerator rack.", "primary_source": "https://nvidianews.nvidia.com/news/nvidia-vera-rubin-platform", "secondary_source": "https://www.hpcwire.com/2026/03/16/nvidia-boasts-7-chips-in-production-for-vera-rubin-platform-including-groq-3-lpu/" }, { "date": "2026-03-17", "date_label": "", "lab": "openai", "type": "model", "title": "GPT-5.4 mini and nano", "summary": "OpenAI released GPT-5.4 mini, available to free users, and GPT-5.4 nano for the API only.", "primary_source": "https://openai.com/index/introducing-gpt-5-4-mini-and-nano/", "secondary_source": "https://www.engadget.com/ai/gpt-54-mini-brings-some-of-the-smarts-of-openais-latest-model-to-chatgpt-free-and-go-users-170000585.html" }, { "date": "2026-03-19", "date_label": "", "lab": "blue-origin", "type": "infrastructure", "title": "Blue Origin files for Project Sunrise", "summary": "Blue Origin filed with the FCC for a space-based computing network of more than 50,000 satellites.", "primary_source": "https://www.techcrunch.com/2026/03/20/jeff-bezos-blue-origin-enters-the-space-data-center-game/", "secondary_source": "https://www.cnbc.com/2026/06/21/do-space-based-ai-data-centers-make-economic-sense.html" }, { "date": "2026-03-24", "date_label": "", "lab": "openai", "type": "retirement", "title": "OpenAI announces end of Sora app and API", "summary": "OpenAI said it would discontinue the Sora consumer app and API to free compute for other products.", "primary_source": "", "secondary_source": "https://engadget.com/ai/openai-is-shutting-down-its-sora-video-generation-app-211023358.html" }, { "date": "2026-03-27", "date_label": "", "lab": "meta", "type": "infrastructure", "title": "Entergy and Meta sign Louisiana power agreement", "summary": "Entergy Louisiana announced seven new gas plants above 5,200 MW for Meta's data center, with about $2 billion in projected customer savings.", "primary_source": "https://www.entergy.com/news/entergy-louisiana-announces-a-new-agreement-with-meta-that-will-deliver-an-additional-2b-in-customer-savings", "secondary_source": "https://www.fortune.com/2026/03/27/meta-hyperion-10-gas-power-plants-louisiana-entergy/" }, { "date": "2026-03-31", "date_label": "", "lab": "openai", "type": "finance", "title": "OpenAI closes $122 billion at an $852 billion valuation", "summary": "OpenAI closed a round of $122 billion in committed capital, including up to $50 billion from Amazon and $30 billion each from Nvidia and SoftBank.", "primary_source": "https://www.cnbc.com/2026/03/31/openai-funding-round-ipo.html", "secondary_source": "https://www.sec.gov/Archives/edgar/data/1018724/000101872426000040/amzn-20260930.htm" }, { "date": "2026-04-01", "date_label": "", "lab": "us-gov", "type": "policy", "title": "American AI Exports Program opens its first call", "summary": "Commerce opened proposals for full-stack AI export packages, through June 30, 2026.", "primary_source": "https://www.trade.gov/press-release/department-commerce-begins-inaugural-call-proposals-american-ai-exports-program", "secondary_source": "" }, { "date": "2026-04-02", "date_label": "", "lab": "google", "type": "model", "title": "Gemma 4 arrives in the Android AICore Developer Preview", "summary": "Google offers Gemma 4 E2B and E4B on Android, and says Gemma 4 is the foundation for the next Gemini Nano.", "primary_source": "https://android-developers.googleblog.com/2026/04/AI-Core-Developer-Preview.html", "secondary_source": "https://developers.google.com/ml-kit/genai" }, { "date": "2026-04-06", "date_label": "", "lab": "broadcom", "type": "chip", "title": "Broadcom discloses long-term TPU deal with Google and 3.5 GW for Anthropic", "summary": "Broadcom's filing describes custom TPU supply for Google and about 3.5 GW of capacity for Anthropic from 2027.", "primary_source": "https://www.sec.gov/Archives/edgar/data/1730168/000119312526144028/d87999d8k.htm", "secondary_source": "https://www.anthropic.com/news/google-broadcom-partnership-compute" }, { "date": "2026-04-07", "date_label": "", "lab": "anthropic", "type": "model", "title": "Claude Mythos Preview (Project Glasswing)", "summary": "Anthropic gave a limited Mythos Preview model to about a dozen organizations for defensive cybersecurity work.", "primary_source": "https://www.anthropic.com/glasswing", "secondary_source": "https://en.wikipedia.org/wiki/Claude_(language_model)" }, { "date": "2026-04-08", "date_label": "", "lab": "meta", "type": "model", "title": "Muse Spark", "summary": "Meta Superintelligence Labs released Muse Spark, its first model, to replace Llama in Meta's assistant.", "primary_source": "https://about.fb.com/news/2026/04/introducing-muse-spark-meta-superintelligence-labs/", "secondary_source": "https://www.theguardian.com/technology/2026/apr/09/meta-first-ai-model-muse-sparks" }, { "date": "2026-04-14", "date_label": "", "lab": "xai", "type": "event", "title": "NAACP sues xAI over Southaven turbines", "summary": "The NAACP sued xAI and MZX Tech under the Clean Air Act over gas turbines allegedly run without a permit.", "primary_source": "https://www.selc.org/news/inside-memphis-fight-against-xai/", "secondary_source": "https://www.cnbc.com/2026/10/31/xai-naacp-clean-air-act-lawsuit-dismissal.html" }, { "date": "2026-04-16", "date_label": "", "lab": "anthropic", "type": "model", "title": "Claude Opus 4.7", "summary": "Anthropic released Claude Opus 4.7.", "primary_source": "https://www.anthropic.com/news/claude-opus-4-7", "secondary_source": "https://www.cnbc.com/2026/04/16/anthropic-claude-opus-4-7-model-mythos.html" }, { "date": "2026-04-17", "date_label": "", "lab": "anthropic", "type": "product", "title": "Claude Design preview", "summary": "Anthropic previewed Claude Design, which turns prompts into visual prototypes, for paid users.", "primary_source": "https://www.anthropic.com/news/claude-design-anthropic-labs", "secondary_source": "https://venturebeat.com/technology/anthropic-just-launched-claude-design-an-ai-tool-that-turns-prompts-into-prototypes-and-challenges-figma" }, { "date": "2026-04-22", "date_label": "", "lab": "google", "type": "chip", "title": "Google announces eighth-generation TPU 8t and TPU 8i", "summary": "Google announced its eighth-generation TPUs as two specialized chips, the TPU 8t for training and the TPU 8i for inference.", "primary_source": "https://blog.google/innovation-and-ai/infrastructure-and-cloud/google-cloud/eighth-generation-tpu-agentic-era/", "secondary_source": "" }, { "date": "2026-04-23", "date_label": "", "lab": "openai", "type": "model", "title": "GPT-5.5 (codename Spud)", "summary": "OpenAI released GPT-5.5 Thinking and Pro, with API access the next day.", "primary_source": "https://openai.com/index/introducing-gpt-5-5/", "secondary_source": "https://www.axios.com/2026/04/23/openai-releases-spud-gpt-model" }, { "date": "2026-04-24", "date_label": "", "lab": "cohere", "type": "event", "title": "Cohere to merge with Aleph Alpha", "summary": "Canada's Cohere announced it would combine with Germany's Aleph Alpha, backed by Schwarz Group, to offer a sovereign alternative to American AI providers.", "primary_source": "https://techcrunch.com/2026/04/25/why-cohere-is-merging-with-aleph-alpha/", "secondary_source": "https://www.cbc.ca/news/business/canada-cohere-ai-investment-9.7401367" }, { "date": "2026-05-05", "date_label": "", "lab": "openai", "type": "model", "title": "GPT-5.5 Instant becomes default", "summary": "GPT-5.5 Instant replaced GPT-5.3 Instant as ChatGPT's default model.", "primary_source": "https://openai.com/index/gpt-5-5-instant/", "secondary_source": "https://en.wikipedia.org/wiki/GPT-5.5" }, { "date": "2026-05-07", "date_label": "", "lab": "openai", "type": "model", "title": "GPT-5.5 with trusted access for cyber", "summary": "OpenAI announced a GPT-5.5 variant for vetted cybersecurity defenders.", "primary_source": "https://openai.com/index/gpt-5-5-with-trusted-access-for-cyber/", "secondary_source": "https://en.wikipedia.org/wiki/GPT-5.5" }, { "date": "2026-05-14", "date_label": "", "lab": "cerebras", "type": "chip", "title": "Cerebras begins trading on Nasdaq", "summary": "Cerebras Class A shares began trading on the Nasdaq under the ticker CBRS, after pricing at 185 US dollars.", "primary_source": "https://www.cerebras.ai/press-release/cerebras-systems-announces-closing-of-initial-public-offering", "secondary_source": "https://www.iposcoop.com/the-ipo-buzz-cerebras-cbrs-prices-ipo-at-185-25-above-range-2/" }, { "date": "2026-05-19", "date_label": "", "lab": "google", "type": "model", "title": "Gemini 3.5 Flash", "summary": "Google launched Gemini 3.5 Flash, positioned around agents.", "primary_source": "https://blog.google/innovation-and-ai/models-and-research/gemini-models/gemini-3-5/", "secondary_source": "https://techcrunch.com/2026/05/19/with-gemini-3-5-flash-google-bets-its-next-ai-wave-on-agents-not-chatbots/" }, { "date": "2026-05-20", "date_label": "", "lab": "xai", "type": "finance", "title": "SpaceX, owner of xAI, files its S-1", "summary": "SpaceX publicly filed to list on Nasdaq, disclosing xAI segment results and $12.7 billion of 2025 AI capital spending.", "primary_source": "https://www.sec.gov/Archives/edgar/data/0001181412/000162828026036936/spaceexplorationtechnologi.htm", "secondary_source": "" }, { "date": "2026-05-28", "date_label": "", "lab": "anthropic", "type": "model", "title": "Claude Opus 4.8", "summary": "Anthropic released Claude Opus 4.8.", "primary_source": "https://www.anthropic.com/news/claude-opus-4-8", "secondary_source": "https://en.wikipedia.org/wiki/Claude_(language_model)" }, { "date": "2026-05-28", "date_label": "", "lab": "anthropic", "type": "finance", "title": "Anthropic raises $65 billion at $965 billion", "summary": "Anthropic announced a $65 billion Series H at a $965 billion post-money valuation, including $15 billion of earlier commitments.", "primary_source": "https://www.anthropic.com/news/series-h", "secondary_source": "https://www.sec.gov/Archives/edgar/data/2130357/000162828026067168/anthropicpbc-sx1.htm" }, { "date": "2026-05-31", "date_label": "", "lab": "nvidia", "type": "chip", "title": "Vera Rubin enters full production", "summary": "Nvidia said Vera Rubin was ramping into full production, with production shipments due to start that fall.", "primary_source": "https://nvidianews.nvidia.com/news/vera-rubin-full-production-agentic-ai-factory", "secondary_source": "" }, { "date": "2026-06-03", "date_label": "", "lab": "eu", "type": "policy", "title": "Commission proposes the Chips Act 2.0", "summary": "The European Commission proposed a Chips Act 2.0 as part of a technological sovereignty package.", "primary_source": "https://digital-strategy.ec.europa.eu/en/library/proposal-chips-act-20", "secondary_source": "https://aeneas-office.org/2026/06/03/european-commission-unveils-proposed-chips-act-2-0/" }, { "date": "2026-06-04", "date_label": "", "lab": "canada", "type": "policy", "title": "Canada's national AI strategy, AI for All", "summary": "Canada launched AI for All, a national AI strategy built on a build-partner-buy approach to sovereign compute, models and adoption.", "primary_source": "https://ised-isde.canada.ca/site/ised/en/node/1046", "secondary_source": "" }, { "date": "2026-06-04", "date_label": "", "lab": "us-states", "type": "policy", "title": "New York legislature passes a data center pause", "summary": "The legislature passed a one-year moratorium on state permits for data centers of 20 MW or more.", "primary_source": "https://www.davispolk.com/insights/client-update/new-york-state-enacts-data-center-moratorium", "secondary_source": "https://www.rockinst.org/blog/updates-on-the-cloud-more-moratoriums-on-data-centers/" }, { "date": "2026-06-04", "date_label": "", "lab": "google", "type": "finance", "title": "Alphabet raises about $50 billion in stock and related securities", "summary": "Alphabet raised about $50 billion net from a public share sale, a Berkshire Hathaway placement and convertible preferred stock, partly for AI infrastructure.", "primary_source": "https://www.sec.gov/Archives/edgar/data/0001652044/000165204426000071/goog-20260630.htm", "secondary_source": "https://www.bankofengland.co.uk/financial-stability-report/2026/july-2026" }, { "date": "2026-06-08", "date_label": "", "lab": "apple", "type": "model", "title": "Apple introduces its third generation of Foundation Models", "summary": "Apple's lineup includes AFM 3 Core at 3 billion parameters and AFM 3 Core Advanced at 20 billion, stored in flash with 1 to 4 billion active.", "primary_source": "https://machinelearning.apple.com/research/introducing-third-generation-of-apple-foundation-models", "secondary_source": "https://thenextweb.com/news/apple-third-generation-foundation-models-afm" }, { "date": "2026-06-09", "date_label": "", "lab": "anthropic", "type": "model", "title": "Claude Fable 5 and Mythos 5", "summary": "Anthropic released Fable 5, a public Mythos-class model with safeguards, and a limited-access Mythos 5.", "primary_source": "https://www.anthropic.com/news/claude-fable-5-mythos-5", "secondary_source": "https://fortune.com/2026/06/09/anthropic-releases-its-first-mythos-model-to-the-public/" }, { "date": "2026-06-11", "date_label": "", "lab": "xai", "type": "finance", "title": "SpaceX prices a record $75 billion IPO", "summary": "SpaceX, which owns xAI, priced a $75 billion IPO at $135 a share; with the underwriters’ full option it sold 638.9 million shares for about $85.7 billion net.", "primary_source": "https://www.sec.gov/Archives/edgar/data/0001181412/000162828026052535/spcx-20260630.htm", "secondary_source": "https://www.iposcoop.com/?p=38639" }, { "date": "2026-06-12", "date_label": "", "lab": "minimax", "type": "model", "title": "MiniMax M3 weights released", "summary": "MiniMax published M3, 428 billion parameters and 23 billion active, under its Community License.", "primary_source": "https://huggingface.co/MiniMaxAI/MiniMax-M3", "secondary_source": "https://x.com/MiniMax_AI/status/2065436935188058208" }, { "date": "2026-06-16", "date_label": "", "lab": "zai", "type": "model", "title": "GLM-5.2", "summary": "Z.ai released GLM-5.2 under MIT with a one-million-token context window.", "primary_source": "https://huggingface.co/zai-org/GLM-5.2", "secondary_source": "https://thenewstack.io/zai-glm-weights-license/" }, { "date": "2026-06-23", "date_label": "", "lab": "china", "type": "infrastructure", "title": "China's LineShine tops the TOP500", "summary": "China's LineShine took first place on the TOP500 at 2.198 exaflops, which Tom's Hardware reports was reached using only CPUs.", "primary_source": "https://top500.org/lists/top500/2026/06/highs/", "secondary_source": "https://www.tomshardware.com/tech-industry/supercomputers/chinas-lineshine-supercomputer-dethrones-us-el-capitan-secures-first-place-in-top-500-list-first-machine-in-the-rankings-to-sustain-more-than-2-exaflops-of-double-precision-performance-using-only-cpus" }, { "date": "2026-06-26", "date_label": "", "lab": "openai", "type": "model", "title": "GPT-5.6 limited preview", "summary": "OpenAI gave GPT-5.6 to a limited group and delayed public rollout after a US government request for early access.", "primary_source": "https://openai.com/index/previewing-gpt-5-6-sol/", "secondary_source": "https://techcrunch.com/2026/06/26/openai-limits-gpt-5-6-rollout-after-government-request-says-restrictions-shouldnt-be-the-norm/" }, { "date": "2026-07-01", "date_label": "", "lab": "anthropic", "type": "finance", "title": "AMD and Anthropic sign a chip and stock agreement", "summary": "AMD committed to buy up to $5 billion of Anthropic stock as GPU milestones are met, and Anthropic expects to spend over $20 billion on AMD systems.", "primary_source": "https://www.sec.gov/Archives/edgar/data/2130357/000162828026067168/anthropicpbc-sx1.htm", "secondary_source": "" }, { "date": "2026-07-09", "date_label": "", "lab": "openai", "type": "retirement", "title": "Atlas browser shutdown announced", "summary": "OpenAI said ChatGPT Atlas would close on 9 August 2026, with its features moving into the ChatGPT desktop app and a Chrome extension.", "primary_source": "https://openai.com/index/chatgpt-for-your-most-ambitious-work/", "secondary_source": "https://www.techradar.com/pro/openai-shuts-down-its-atlas-browser-after-not-even-a-year" }, { "date": "2026-07-09", "date_label": "", "lab": "openai", "type": "model", "title": "GPT-5.6 (Luna, Terra, Sol) and ChatGPT Work", "summary": "OpenAI publicly released the GPT-5.6 family in three tiers and launched ChatGPT Work, an agent combining ChatGPT and Codex for office tasks.", "primary_source": "https://openai.com/index/gpt-5-6/", "secondary_source": "https://techcrunch.com/2026/07/09/openai-launches-its-new-family-of-models-with-gpt-5-6/" }, { "date": "2026-07-13", "date_label": "", "lab": "meta", "type": "infrastructure", "title": "Meta expands its Louisiana site to 5 GW", "summary": "Meta said its Richland Parish site, home of the Hyperion cluster, will expand to 5 GW with more than $50 billion invested.", "primary_source": "https://datacenters.atmeta.com/2026/07/deepening-our-investment-in-richland-parish-louisiana/", "secondary_source": "" }, { "date": "2026-07-14", "date_label": "", "lab": "us-states", "type": "policy", "title": "New York imposes the first state data center moratorium", "summary": "Governor Hochul's Executive Order 62 paused state permitting for data centers of 50 MW or more during an environmental review.", "primary_source": "https://www.davispolk.com/insights/client-update/new-york-state-enacts-data-center-moratorium", "secondary_source": "https://rockinst.org/blog/pushing-back-on-the-cloud-the-first-statewide-moratorium/" }, { "date": "2026-07-15", "date_label": "", "lab": "thinking-machines", "type": "model", "title": "Thinking Machines releases Inkling", "summary": "Thinking Machines Lab released Inkling, its first general model, with 975 billion parameters and 41 billion active, under Apache 2.0.", "primary_source": "https://huggingface.co/thinkingmachines/Inkling", "secondary_source": "https://aiweekly.co/alerts/poolside-thinking-machines-moonshot-ship-open-weight-models" }, { "date": "2026-07-16", "date_label": "", "lab": "moonshot", "type": "model", "title": "Kimi K3 launches", "summary": "Moonshot launched Kimi K3 as a hosted model of 2.8 trillion parameters, promising the weights by July 27.", "primary_source": "https://www.kimi.com/blog/kimi-k3", "secondary_source": "https://thenewstack.io/kimi-k3-open-weights/" }, { "date": "2026-07-16", "date_label": "", "lab": "china", "type": "policy", "title": "China-backed WAICO is established", "summary": "CSIS reports that representatives of 29 countries signed an agreement in Shanghai creating the World Artificial Intelligence Cooperation Organization.", "primary_source": "https://www.csis.org/analysis/state-ai-global-governance-and-its-implications-us-china-summit", "secondary_source": "https://un.china-mission.gov.cn/eng/zgyw/202607/t20260717_11984766.htm" }, { "date": "2026-07-22", "date_label": "", "lab": "amd", "type": "chip", "title": "AMD and Anthropic sign a deal for up to 2 GW of MI450 GPUs", "summary": "Anthropic will deploy up to 2 GW of AMD MI450 GPUs from 2027, and AMD will invest up to $5 billion in Anthropic.", "primary_source": "https://ir.amd.com/news-events/press-releases/detail/1292/amd-and-anthropic-announce-strategic-partnership-to-deploy-up-to-2-gigawatts-of-amd-instinct-mi450-series-gpus", "secondary_source": "" }, { "date": "2026-07-23", "date_label": "", "lab": "amd", "type": "hardware", "title": "AMD launches MI455X and Helios", "summary": "At Advancing AI 2026, AMD said its Helios rack, with 72 MI455X GPUs, was in production, and OpenAI said it expected to start bringing Helios online in the fourth quarter of 2026.", "primary_source": "https://ir.amd.com/news-events/press-releases/detail/1294/aai-2026-amd-delivers-full-stack-compute-for-the-agentic-ai-era", "secondary_source": "https://www.itpro.com/infrastructure/amd-advancing-ai-2026-helios-on-the-rise-with-launch-of-new-instinct-gpus" }, { "date": "2026-07-24", "date_label": "", "lab": "anthropic", "type": "model", "title": "Claude Opus 5", "summary": "Anthropic released Opus 5, close to Fable 5 capability at about half the price.", "primary_source": "https://www.anthropic.com/news/claude-opus-5", "secondary_source": "https://en.wikipedia.org/wiki/Claude_(language_model)" }, { "date": "2026-07-24", "date_label": "", "lab": "swiss-ai", "type": "model", "title": "Apertus 1.5", "summary": "The Swiss AI Initiative released Apertus 1.5, adding images, optional reasoning and a 262,144-token context window to its fully open 8B and 70B models.", "primary_source": "https://publicai.co/stories/apertus-1-5", "secondary_source": "" }, { "date": "2026-07-27", "date_label": "", "lab": "moonshot", "type": "model", "title": "Kimi K3 weights released", "summary": "Moonshot published the 1.56 TB Kimi K3 weights on Hugging Face under the custom Kimi K3 License.", "primary_source": "https://huggingface.co/moonshotai/Kimi-K3", "secondary_source": "https://thenewstack.io/kimi-k3-open-weights/" }, { "date": "2026-07-30", "date_label": "", "lab": "eu", "type": "policy", "title": "EuroHPC launches the AI Gigafactories call", "summary": "EuroHPC opened a tender for up to seven AI gigafactories, with bids due November 12, 2026.", "primary_source": "https://www.eurohpc-ju.europa.eu/eurohpc-joint-undertaking-launches-ai-gigafactories-call-2026-07-30_en", "secondary_source": "https://www.sullcrom.com/insights/memo/2026/October/EuroHPC-Joint-Undertaking-Launches-AI-Gigafactories-Call" }, { "date": "2026-08", "date_label": "", "lab": "zai", "type": "model", "title": "GLM-5.3-Flash", "summary": "Z.ai released GLM-5.3-Flash, a new natively multimodal base of 320 billion parameters and 18 billion active, under MIT.", "primary_source": "https://huggingface.co/zai-org/GLM-5.3-Flash", "secondary_source": "https://thenewstack.io/zai-glm-weights-license/" }, { "date": "2026-08-01", "date_label": "", "lab": "nvidia", "type": "finance", "title": "Nvidia guarantees up to $105 billion for an OpenAI campus", "summary": "Nvidia agreed to guarantees capped at $105 billion on leases for about 4.25 gigawatts in Ohio, tied to an OpenAI affiliate.", "primary_source": "https://www.sec.gov/Archives/edgar/data/1045810/000104581026000075/nvda-20260726.htm", "secondary_source": "" }, { "date": "2026-08-11", "date_label": "", "lab": "nvidia", "type": "model", "title": "Nemotron 3.5 Lightning", "summary": "NVIDIA released Nemotron 3.5 Lightning, 30 billion parameters and 3 billion active, under OpenMDW 1.1.", "primary_source": "https://huggingface.co/nvidia/NVIDIA-Nemotron-3.5-Lightning-30B-A3B-NVFP4", "secondary_source": null }, { "date": "2026-08-14", "date_label": "", "lab": "alibaba", "type": "model", "title": "Qwen3.8-27B", "summary": "Alibaba released Qwen3.8-27B, a dense vision-language model, under Apache 2.0.", "primary_source": "https://huggingface.co/Qwen/Qwen3.8-27B", "secondary_source": "https://en.wikipedia.org/wiki/Qwen" }, { "date": "2026-08-21", "date_label": "", "lab": "deepseek", "type": "model", "title": "DeepSeek-V4-Flash-Vision-Exp", "summary": "DeepSeek added multimodal input to its API with V4-Flash-Vision-Exp.", "primary_source": "https://api-docs.deepseek.com/news/news260910", "secondary_source": "https://huggingface.co/deepseek-ai/DeepSeek-V4-Flash-Vision-Exp" }, { "date": "2026-08-25", "date_label": "", "lab": "apple", "type": "hardware", "title": "Apple announces the Mac Studio with M5 Max and M5 Ultra", "summary": "The new Mac Studio offers up to 512 GB of unified memory and 1.2 TB/s of memory bandwidth on the M5 Ultra.", "primary_source": "https://www.apple.com/mac-studio/specs/", "secondary_source": "https://9to5mac.com/2026/08/25/apple-unveils-next-generation-mac-studio-with-m5-max-and-m5-ultra/" }, { "date": "2026-08-26", "date_label": "", "lab": "alibaba", "type": "model", "title": "Qwen3.8-Flash-Next", "summary": "Alibaba published Qwen3.8-Flash-Next, a multimodal mixture-of-experts model of 180 billion parameters, under the Qwen Community License.", "primary_source": "https://huggingface.co/Qwen/Qwen3.8-Flash-Next", "secondary_source": "https://en.wikipedia.org/wiki/Qwen" }, { "date": "2026-08-28", "date_label": "", "lab": "zai", "type": "model", "title": "GLM-5.3 weights released", "summary": "Z.ai published GLM-5.3 under a custom licence.", "primary_source": "https://huggingface.co/zai-org/GLM-5.3", "secondary_source": "https://thenewstack.io/zai-glm-weights-license/" }, { "date": "2026-09", "date_label": "", "lab": "ai-singapore", "type": "model", "title": "SEA-LION v4.8", "summary": "AI Singapore released SEA-LION v4.8, with models built on NVIDIA Nemotron-3 in 30B and 120B sizes.", "primary_source": "https://docs.sea-lion.ai/models/sea-lion-v4.8", "secondary_source": "https://huggingface.co/aisingapore/Nemotron-SEA-LION-v4.8-30B-A3B" }, { "date": "2026-09-01", "date_label": "", "lab": "nsf", "type": "policy", "title": "NSF funds the NAIRR Operations Center", "summary": "The NSF awarded the San Diego Supercomputer Center a five-year, $35 million award, with TACC as partner, to move NAIRR from pilot to long-term capability.", "primary_source": "https://www.sdsc.edu/news/2026/PR20260901-NAIRR-OC.html", "secondary_source": "https://www.nextgov.com/artificial-intelligence/2026/09/nsf-launches-new-ai-research-operations-center/415786/" }, { "date": "2026-09-03", "date_label": "", "lab": "openai", "type": "model", "title": "GPT-6 Astra", "summary": "OpenAI unveiled GPT-6 Astra, a model strong at computer use, for approved testers, with a restricted version for paid ChatGPT users the next day.", "primary_source": "https://openai.com/index/gpt-6-astra/", "secondary_source": "https://www.axios.com/2026/09/03/openai-astra-gpt-6-agi-brockman" }, { "date": "2026-09-08", "date_label": "", "lab": "meta", "type": "product", "title": "Meta launches the Muse personal agent", "summary": "Meta launched Muse, a personal AI agent that runs on a dedicated cloud virtual machine, in the US, free for most uses with subscription plans.", "primary_source": "https://about.fb.com/news/2026/09/introducing-muse-personal-ai-agent/", "secondary_source": "https://techcrunch.com/2026/09/08/meta-debuts-its-muse-ai-agent-will-consumers-trust-it/" }, { "date": "2026-09-10", "date_label": "", "lab": "deepseek", "type": "model", "title": "DeepSeek-V4.1-Flash", "summary": "DeepSeek released V4.1-Flash, with 552 billion parameters, 8 and 16 billion active, and began phasing out V4-Pro.", "primary_source": "https://api-docs.deepseek.com/news/news260910", "secondary_source": "https://huggingface.co/deepseek-ai/DeepSeek-V4.1-Flash" }, { "date": "2026-09-14", "date_label": "", "lab": "bell", "type": "infrastructure", "title": "Bell expands its Saskatchewan AI plan", "summary": "Saskatchewan and Bell announced a path to 1.2 GW and more than $50 billion in estimated investment.", "primary_source": "https://www.saskatchewan.ca/government/news-and-media/2026/september/14/single-largest-private-sector-capital-investment-in-history-of-canada-through-bell-ais-saskatchewan", "secondary_source": "https://canada.constructconnect.com/joc/news/projects/2026/10/construction-of-bell-canadas-ai-data-centre-in-saskatchewan-scaling-up" }, { "date": "2026-09-16", "date_label": "", "lab": "cohere", "type": "event", "title": "Cohere and Aleph Alpha sign a definitive merger agreement", "summary": "The companies signed a definitive agreement, with dual headquarters in Toronto and Berlin, still subject to regulatory approvals.", "primary_source": "https://live.euronext.com/en/financial-news/cohere-aleph-alpha-combine-target-enterprise-ai-market", "secondary_source": "" }, { "date": "2026-09-21", "date_label": "", "lab": "xiaomi", "type": "model", "title": "MiMo-V2.6 series", "summary": "Xiaomi published MiMo-V2.6-Pro and Flash under MIT.", "primary_source": "https://huggingface.co/XiaomiMiMo/MiMo-V2.6-Pro-RL", "secondary_source": "https://www.unite.ai/xiaomis-new-flagship-model-leads-open-weight-rankings-with-a-score-of-46/" }, { "date": "2026-09-28", "date_label": "", "lab": "anthropic", "type": "model", "title": "Claude Sonnet 5.5", "summary": "Anthropic released Sonnet 5.5, a faster lower-cost companion to Opus 5.5.", "primary_source": "https://www.anthropic.com/claude-sonnet-5-5", "secondary_source": "https://en.wikipedia.org/wiki/Claude_(language_model)" }, { "date": "2026-09-29", "date_label": "", "lab": "openai", "type": "event", "title": "DevDay 2026 and GPT-6.1 Sol", "summary": "At DevDay in San Francisco OpenAI launched GPT-6.1 Sol, near Astra-level at a fifth of the price, and always-on agents called Dots; GPT-6.1 Astra was held back.", "primary_source": "https://openai.com/index/introducing-gpt-6-1-sol/", "secondary_source": "https://techcrunch.com/2026/09/29/openai-launches-gpt-6-1-sol-says-it-nearly-matches-gpt-6-astra-and-costs-less/" }, { "date": "2026-09-29", "date_label": "", "lab": "openai", "type": "product", "title": "OpenAI launches dots, always-on agents", "summary": "OpenAI announced dots at DevDay, personal agents powered by GPT-6 Astra, for ChatGPT Pro and Business Premium users in eligible markets.", "primary_source": "https://techcrunch.com/2026/09/29/openai-launches-dots-its-bubbly-agentic-avatar/", "secondary_source": "https://www.aljazeera.com/economy/2026/9/30/openai-launches-dots-personal-ai-assistant-built-to-handle-everything" }, { "date": "2026-10", "date_label": "", "lab": "moonshot", "type": "model", "title": "Kimi K3.1 Pro", "summary": "Moonshot launched K3.1 Pro, a further post-trained K3, on its services and promised the weights by November 4.", "primary_source": "https://www.kimi.com/blog/kimi-k3-1-pro", "secondary_source": "https://braindetox.kr/en/posts/kimi_k3_1_pro_release_2026.html" }, { "date": "2026-10-01", "date_label": "", "lab": "cloudflare", "type": "model", "title": "Cloudflare's Clef decision models", "summary": "Cloudflare released Clef and Clef-flash, decision models post-trained from Qwen3.8-27B that return typed answers with probabilities, under Apache 2.0.", "primary_source": "https://blog.cloudflare.com/clef-decision-models", "secondary_source": "https://huggingface.co/Cloudflare/clef" }, { "date": "2026-10-03", "date_label": "", "lab": "aleph-alpha", "type": "model", "title": "Aleph Alpha releases Kolibri-1", "summary": "Aleph Alpha released Kolibri, an English-German model of 78 billion parameters, under Apache 2.0.", "primary_source": "https://huggingface.co/Aleph-Alpha/Kolibri-1", "secondary_source": "https://www.trendingtopics.eu/aleph-alpha-kolibri-open-weight/" }, { "date": "2026-10-05", "date_label": "", "lab": "reflection", "type": "model", "title": "Reflection announces Beam", "summary": "Reflection AI unveiled Beam, its first open-weight model, with 501 billion parameters and 23 billion active.", "primary_source": "https://reflection.ai/blog/introducing-beam", "secondary_source": "https://techcrunch.com/2026/10/05/reflection-debuts-beam-a-open-weight-ai-model-to-rival-chinese-models-at-lower-compute-cost/" }, { "date": "2026-10-07", "date_label": "", "lab": "openai", "type": "model", "title": "GPT-6 Instant and Intelligent UI for everyone", "summary": "OpenAI brought GPT-6 to all ChatGPT tiers with interactive 'Intelligent UI' answers, paid users first and Free and Go users a day later.", "primary_source": "https://openai.com/index/gpt-6-for-everyone/", "secondary_source": "https://betanews.com/article/gpt-6-intelligent-ui-chatgpt-free-users/" }, { "date": "2026-10-07", "date_label": "", "lab": "anthropic", "type": "model", "title": "Claude Haiku 5.5", "summary": "Anthropic released Haiku 5.5, a small low-cost model for high-volume tasks, and halved Sonnet 5.5's cache-read price.", "primary_source": "https://www.anthropic.com/claude-haiku-5-5", "secondary_source": "https://en.wikipedia.org/wiki/Claude_(language_model)" }, { "date": "2026-10-08", "date_label": "", "lab": "anthropic", "type": "policy", "title": "Anthropic updates its Usage Policy", "summary": "Anthropic published a new Usage Policy, effective November 12, banning sustained needless cruelty toward its models and removing its blanket ban on personalized political targeting.", "primary_source": "https://www.anthropic.com/news/2026-usage-policy-update", "secondary_source": "https://www.theregister.com/ai-and-ml/2026/10/09/anthropic-asks-users-to-stop-being-mean-to-claude/5302218" }, { "date": "2026-10-12", "date_label": "", "lab": "meta", "type": "model", "title": "Muse Wonder and Muse Wonder Full", "summary": "Meta released Muse Wonder, its first frontier-class model, via its API and Muse Code, with Wonder Full for subscribers.", "primary_source": "https://research.meta.ai/blog/muse-wonder", "secondary_source": "https://aimagazine.com/news/muse-wonder-how-is-meta-competing-at-the-frontier" }, { "date": "2026-10-12", "date_label": "", "lab": "intel", "type": "chip", "title": "Intel GPU D-series (Crescent Island)", "summary": "Intel shared details of its GPU D-series for inference, previously codenamed Crescent Island.", "primary_source": "https://www.intc.com/news-events/press-releases/detail/1783/intel-accelerates-open-infrastructure-for-the-agentic-ai-era", "secondary_source": "" }, { "date": "2026-10-15", "date_label": "", "lab": "tsmc", "type": "finance", "title": "TSMC reports $46.72 billion quarter, with 68% from high-performance computing", "summary": "TSMC reported third-quarter revenue up 41.2% in US dollars.", "primary_source": "https://www.sec.gov/Archives/edgar/data/0001046179/000104617926000726/a3q26e_withguidancexfinal.htm", "secondary_source": "https://investor.tsmc.com/english/encrypt/files/encrypt_file/reports/2026-10/eba4fac72873bf546b669bd2b5d973df5c93215d/TSMC%203Q26%20Transcript.pdf" }, { "date": "2026-10-15", "date_label": "", "lab": "nvidia", "type": "model", "title": "Nemotron 3.5 Super VL and Bolt 74B", "summary": "NVIDIA released Super VL, a 120 billion parameter vision model.", "primary_source": "https://huggingface.co/nvidia/NVIDIA-Nemotron-3.5-Super-VL-120B-A12B-NVFP4", "secondary_source": "https://blogs.nvidia.com/blog/nemotron-3-5/" }, { "date": "2026-10-15", "date_label": "", "lab": "meta", "type": "model", "title": "Muse Spark Open weights", "summary": "Meta released Muse Spark Open, 816 billion parameters and 42 billion active, under the Muse Model License.", "primary_source": "https://research.meta.ai/blog/muse-spark-open", "secondary_source": "https://pytorch.org/blog/muse-spark-open/" }, { "date": "2026-10-16", "date_label": "", "lab": "reflection", "type": "model", "title": "Beam weights released", "summary": "Reflection published the Beam weights on Hugging Face under Apache 2.0, with over 40 partners and integrations.", "primary_source": "https://reflection.ai/blog/beam-is-available", "secondary_source": "https://huggingface.co/reflection/Beam-1.0-501B-A23B" }, { "date": "2026-10-17", "date_label": "", "lab": "deepseek", "type": "model", "title": "DeepSeek-V4.1-Flash-1017", "summary": "DeepSeek published a minor update to V4.1-Flash reporting 77.0 on DeepSWE v1.1, up from 74.2.", "primary_source": "https://huggingface.co/deepseek-ai/DeepSeek-V4.1-Flash-1017", "secondary_source": "https://api-docs.deepseek.com/news/news260910" }, { "date": "2026-10-21", "date_label": "", "lab": "anthropic", "type": "product", "title": "Anthropic previews Claude Duet", "summary": "Anthropic opened a waitlisted US research preview of Claude Duet, a proactive agent built on Claude Opus 5.5 for professionals and small business owners.", "primary_source": "https://claude.com/resources/articles/claude-duet", "secondary_source": "https://www.cnbc.com/2026/10/21/anthropic-preview-of-claude-duet-personal-ai-agent-take-on-muse-dots.html" }, { "date": "2026-10-23", "date_label": "", "lab": "nvidia", "type": "hardware", "title": "NVIDIA's 64 GB DGX Spark goes on sale through partners", "summary": "A 64 GB DGX Spark goes on sale from Acer, ASUS, Dell, Gigabyte, HP and MSI, starting at US$4,999.", "primary_source": "https://blogs.nvidia.com/blog/local-ai-dgx-spark-64gb-sync/", "secondary_source": "https://www.implicator.ai/nvidia-adds-64gb-dgx-spark-at-4-999-as-128gb-model-climbs-to-6-950/" }, { "date": "2026-10-23", "date_label": "", "lab": "minimax", "type": "model", "title": "MiniMax M3.1-Flash weights", "summary": "MiniMax published M3.1-Flash with sparse attention throughout and a 252 GB download.", "primary_source": "https://huggingface.co/MiniMaxAI/MiniMax-M3.1-Flash", "secondary_source": null }, { "date": "2026-10-23", "date_label": "", "lab": "nvidia", "type": "model", "title": "Nemotron 3.5 Bolt E4B", "summary": "NVIDIA released Bolt E4B, post-trained from Google's Gemma 4 E4B, with a 4 GB version for edge devices.", "primary_source": "https://huggingface.co/nvidia/NVIDIA-Nemotron-3.5-Bolt-E4B-BF16", "secondary_source": "https://blogs.nvidia.com/blog/nemotron-3-5/" }, { "date": "2026-10-26", "date_label": "", "lab": "anthropic", "type": "finance", "title": "Anthropic files publicly for an IPO", "summary": "Anthropic filed a Form S-1 with the SEC to list Class A shares on the Nasdaq under the symbol ANT, without a price or share count.", "primary_source": "https://www.sec.gov/Archives/edgar/data/2130357/000162828026067168/anthropicpbc-sx1.htm", "secondary_source": "https://techcrunch.com/2026/10/26/anthropic-officially-files-for-its-public-debut/" }, { "date": "2026-10-26", "date_label": "", "lab": "openai", "type": "model", "title": "GPT-6.1 Terra", "summary": "OpenAI released GPT-6.1 Terra, a fast model for computer use, vision and office work at $1 and $5 per million tokens, rated Critical for cybersecurity.", "primary_source": "https://openai.com/index/introducing-gpt-6-1-terra/", "secondary_source": "https://www.infoworld.com/article/4241006/openai-bets-on-ai-economics-with-lower-cost-gpt-6-1-terra.html" }, { "date": "2026-10-28", "date_label": "", "lab": "sk-hynix", "type": "finance", "title": "SK hynix posts record quarter and expects memory shortage through 2030", "summary": "SK hynix reported 104.6 trillion won in revenue and said more than half of 2027 and 2028 capacity is under long-term contracts.", "primary_source": "https://www.koreatimes.co.kr/business/companies/20261028/sk-hynix-logs-record-q3-operating-profit-of-628-bil", "secondary_source": "https://www.trendforce.com/news/?p=65818" }, { "date": "2026-10-28", "date_label": "", "lab": "google", "type": "model", "title": "Gemini 4 Carbon", "summary": "Google announced Gemini 4 Carbon, its strongest model, available first to a subset of cyber defenders through the Fairwind Program.", "primary_source": "https://blog.google/innovation-and-ai/models-and-research/gemini-models/introducing-gemini-4-carbon/", "secondary_source": "https://runtimewire.com/article/google-gemini-4-carbon-cybersecurity-launch" }, { "date": "2026-10-29", "date_label": "", "lab": "canada", "type": "policy", "title": "Canada takes a stake in Cohere", "summary": "Prime Minister Mark Carney announced an equity investment of up to $1.6 billion in Cohere, tied to jobs, compute in Canada and keeping its headquarters here.", "primary_source": "https://www.newswire.ca/news-releases/canada-announces-landmark-investment-in-cohere-to-scale-secure-sovereign-ai-814597810.html", "secondary_source": "https://www.cbc.ca/news/business/canada-cohere-ai-investment-9.7401367" }, { "date": "2026-10-29", "date_label": "", "lab": "mistral", "type": "model", "title": "Mistral Large 4 weights released", "summary": "Mistral published the improved Large 4 weights under a modified MIT licence that requires a commercial licence for model-as-a-service businesses.", "primary_source": "https://mistral.ai/news/mistral-large-4-oss/", "secondary_source": "https://huggingface.co/mistralai/Mistral-Large-4-1T-A52B" }, { "date": "2026-10-30", "date_label": "", "lab": "markets", "type": "finance", "title": "AI spending worries drive a tech sell-off", "summary": "The Nasdaq fell 2.9%, its largest one-day percentage drop since June 5, after results from Alphabet, Amazon and Intel raised concern about AI spending.", "primary_source": "https://finance.yahoo.com/markets/stocks/articles/nasdaq-futures-sink-amazon-intel-114737736.html", "secondary_source": "https://www.sec.gov/Archives/edgar/data/1652044/000165204426000080/googexhibit991q32026.htm" }, { "date": "2026-11-02", "date_label": "", "lab": "anthropic", "type": "finance", "title": "Anthropic's S-1 amendment adds Q3 results and $895 billion in commitments", "summary": "Amendment No. 1 reports a $567 million third-quarter operating loss and about $895.4 billion of compute and infrastructure commitments; the price remains blank.", "primary_source": "https://www.sec.gov/Archives/edgar/data/2130357/000162828026068676/anthropicpbc-sx1ax1.htm", "secondary_source": "" } ] }