The Omniscient Bulletin · November 12, 2026
Huang quits Nvidia as a chip crackdown stalls at Commerce
Nvidia's board went from backing Jensen Huang on Sunday to meeting late Wednesday on removing him, and on Thursday he resigned as CEO, days before the company reports results on November 17. Washington looks no more decisive, with Axios reporting that the Commerce Department's chief has repeatedly held up a chip-diversion crackdown that Trump approved. On AI agents that shop for people, United and Brussels reached opposite answers: one is banning them, the other proposed giving their users clear rights.
Industry
Jensen Huang resigned as Nvidia CEO after a board meeting on ousting him and a report of a second affair, this one with an executive who reported to him
Nvidia said Thursday that Huang resigned as CEO, president and director, effective immediately; CFO Colette Kress is interim CEO and Harvey Jones the board's first chair. It followed a board meeting late Wednesday, Reuters reported, and a New York Times report, which Reuters could not verify, of a second affair, with Kimberly Powell, an executive who reported to him. Bloomberg said the board had used an outside firm to find successors. Results are still due Nov. 17.
Policy
Trump has approved a crackdown on companies suspected of diverting Nvidia chips to China three times this year, but Commerce Secretary Lutnick keeps stalling it, Axios reports
President Trump last Thursday ordered action against companies suspected of funneling Nvidia chips to China, the third such order this year, after consulting Elon Musk, but Commerce Secretary Lutnick has repeatedly blocked it, Axios reported Wednesday. In the Situation Room last Thursday, Treasury Secretary Scott Bessent confronted Lutnick over his department's resistance to export controls; Lutnick seemed to agree but said he had to check with Trump. Reuters noted the US has added no names to its Entity List since October 2025.
Industry
Anthropic's IPO has drawn orders for several times the shares on offer, Bloomberg reports, as it and its co-founders seek up to $90.4 billion
Anthropic's IPO is oversubscribed multiple times, Bloomberg reported, citing people familiar with the matter, with demand led by anchor-size indications of interest. Anthropic and its co-founders seek up to $90.4 billion from the IPO and a private placement, which Bloomberg said could be the largest first-time share sale ever. Tuesday's amended filing set a $70 to $78 range for about 801.4 million new shares, plus $26 billion of nonvoting Class N stock for seven tech firms. No final price or trading date has been announced.
Research
Google DeepMind released AlphaGenome 2, which predicts how DNA variants act in specific human cell types
Google DeepMind released AlphaGenome 2, which predicts how DNA variants affect gene activity in specific human cell types and states rather than tissue averages. Trained on over 100 million single-cell profiles from 68 tissues, it flags nearly twice as many expression-altering variants as its predecessor; on an immune-cell dataset it found 30% of causal variants vs. 16% for the original model fine-tuned on single-cell data. It adds DNA methylation and handles new cell states without retraining. Non-commercial research access is via API.
Compute
Britain's Sovereign AI fund and the British Business Bank backed chip startup Fractile in a round of over 500 million pounds that values it at over 5 billion pounds
The British Business Bank and the UK government's Sovereign AI Venture Fund have invested in Fractile, a London and Bristol startup designing AI chips, in a Series C round worth over 500 million pounds that values the company at over 5 billion pounds, the government said Wednesday. CapitalG, Lightspeed and Redpoint also backed the round. Fractile is rebuilding AI hardware to be faster, cheaper and more energy efficient, working across chips, software and machine learning research, and plans to expand in Bristol.
Policy
The European Commission proposed a Digital Fairness Act that would give shoppers clear rights when an AI agent buys on their behalf
The European Commission proposed a Digital Fairness Act giving consumers "clear rights when using an AI agent to buy products on their behalf" and a right to reach a human in customer service. It would ban manipulative "dark patterns" such as fake claims of near sell-outs, and make subscriptions easier to cancel. A companion revision of enforcement rules would let the Commission investigate large cross-border cases itself, with fines of up to 6% of worldwide annual turnover. Parliament and Council now take them up.
Agents
United Airlines is banning third-party personal AI agents from booking its flights, citing security risks and federal rules
United said Wednesday it is banning third-party personal AI agents from booking flights for travelers, citing security risks and possible breaches of federal rules. It said sharing account credentials with them can expose sensitive information and lead to "unintended purchases, account changes, or use of MileagePlus miles," and that agent check-in could breach rules making passengers confirm in person their bags hold no banned substances. United will restrict access on its platforms and via travel agencies and aggregators.