A boardroom crisis at the world’s most valuable chipmaker, the biggest listing in history, a DNA model that reads cells one by one, and an airline telling AI agents to look but not book.
Nvidia’s Board Prepares to Remove Jensen Huang
The Wall Street Journal reported on Wednesday that Nvidia’s board is preparing to remove Jensen Huang as chief executive, after reports of two extramarital affairs, one of them with a woman American intelligence officials suspect of spying for China. The turn is stunning because just three days earlier, the same board opened an independent investigation and publicly expressed confidence in Huang. According to the Journal, directors were set to meet as soon as Wednesday to decide his future, and no final decision had been announced by Thursday morning.
The sequence unfolded fast. On Sunday, Huang told employees in a letter that he had been unfaithful to his wife. Then on Wednesday, the New York Times reported a second alleged affair, this one with Kimberly Powell, Nvidia’s vice president of healthcare, who reports directly to him.
The market reacted immediately. Nvidia shares fell about nine percent in after-hours trading on Wednesday. Across the two sessions since Huang’s admission, the stock is down almost ten percent, wiping out more than four hundred billion dollars in market value and leaving the shares about thirty percent below their early October record.
A Succession Problem With No Plan
The practical problem is succession. Nvidia has no chief operating officer, no board chair, and no public succession plan, and more than fifty managers report directly to Huang, who co-founded the company in 1993. Whoever takes over — if the board follows through — inherits a company whose entire operating structure runs through one person.
Anthropic’s Record IPO Keeps Getting Bigger
From one AI giant in turmoil to another AI giant about to make history: Anthropic’s IPO is already multiple times oversubscribed, according to Bloomberg, just two days into the roadshow. The demand is mostly coming from anchor-size orders, meaning big institutions asking for big blocks of shares.
Anthropic is selling eight hundred twenty-five million shares at seventy to seventy-eight dollars each, which would raise up to about sixty-four billion dollars, and a twenty-six billion dollar private placement to chip and cloud companies takes the combined total to as much as ninety point four billion dollars. That would make it the largest stock market listing ever, topping SpaceX. The Financial Times has reported that underwriters are discussing pricing the deal above the seventy-eight dollar top of the range, which would push the fully diluted valuation beyond roughly one point four trillion dollars. Pricing is set for November 18, with trading starting November 19. One caution, though: indications of interest are not binding.
AlphaGenome 2 Reads DNA Cell by Cell
Google DeepMind announced something genuinely new on Tuesday: AlphaGenome 2, a model that predicts how DNA variants affect gene activity in specific human cell types, not just in broad tissue averages. That distinction matters. A genetic variant can be harmless in most of the body but dangerous in one particular kind of cell, and tissue-wide averages wash that signal out.
AlphaGenome 2 was trained on more than one hundred million individual cell profiles across sixty-eight tissues, and DeepMind says it identifies nearly twice as many variants that change gene expression as its previous model. The proof point that caught scientists’ attention is a real one: researchers used the model to replicate aspects of a 2024 discovery about inflammatory bowel disease, tracing it to a gene called ETS2 that is overactive in immune cells.
AlphaGenome 2 is available now for non-commercial research through the AlphaGenome API, with commercial research in private preview. One important caveat: it is for research only, and it has not been validated or approved for clinical or diagnostic use.
United Airlines to AI Agents: Look, Don’t Book
United Airlines announced on Wednesday that it will block unauthorized third-party AI assistants from booking, changing, or paying for its flights. United is the first major American airline to draw this line. The ban covers its website and app, and also travel agencies and aggregators, and United updated its contract of carriage so it can cancel any reservation made or changed by an AI agent. But here is the nuance: assistants can still show you United flights and fares. They just cannot complete the purchase.
United gave three reasons. Sharing account credentials with a third-party assistant can expose sensitive information. An agent could make unintended purchases, change a booking, or spend your MileagePlus miles. And automated check-in could break federal rules, because a traveler has to personally attest that their checked bags hold no prohibited items, like explosives.
United chief executive Scott Kirby put it bluntly at an event on Tuesday: “To be honest, I do not want to create another middleman.” But not every airline agrees. JetBlue president Marty St. George said on Wednesday that airlines should follow their customers’ lead on AI booking agents.
The Fed Warns on AI Debt
A warning from the Federal Reserve: Governor Michelle Bowman told the Council on Foreign Relations on Tuesday that AI-related debt is on track to exceed eight hundred billion dollars this year, and that a large correction in stock valuations is among the top risks to the economy if AI returns fall short of expectations. She also noted that leases and contract commitments tied to the buildout could add potential payments of more than three times that debt.
Quick Hits: Gemini Decisions and ChatGPT’s One Voice
Two more developments rounded out the day:
- Google launched the Gemini Decisions API on Tuesday. Built on a new model called Gemini Decision 1, it scores your options instead of writing text: one request can evaluate up to one thousand yes-or-no, multiple-choice, or rubric questions against a single input, across text, images, audio, video, and PDFs, in about one hundred fifty to two hundred fifty milliseconds. It costs ten cents per million input tokens, with no charge for output tokens, and there is a free tier in Google AI Studio.
- OpenAI is retiring two ChatGPT voice modes. Advanced and Standard Voice disappear from the settings menu on December 11, leaving GPT-Live as the only voice option. OpenAI says voice usage has nearly doubled since GPT-Live launched in July. Live video and screen sharing will keep running on Advanced Voice for now, because GPT-Live cannot do those yet.
Sources
- Nvidia board preparing to remove CEO Jensen Huang: Reuters
- United Airlines bans AI agents from booking flights; JetBlue takes opposite stance: Reuters
- Fed Governor Michelle Bowman warns AI debt could trigger market correction: Bloomberg Law
Adapted from the November 12, 2026 episode of Daily Tech and AI.
Daily Tech and AI is AI-voiced and produced by Angelo Ruiz.
Leave a Reply