Jump within the current audio brief. Story thumbnails mirror the cards below when available.
Same-day briefs for Fri Nov 13, sorted by release time. The selected date stays fixed.
Changing the date preserves the current brief type: AI News PM.
/ai-news-pm/2026-11-13/.Story cards stay below the sticky dock while audio, chapters, date, and brief navigation remain accessible.
Anthropic's offering drew roughly six times the orders available, seeking up to $64.4 billion at a $1.4 trillion valuation, with a concurrent $26 billion private placement pushing the combined raise toward $90.4 billion — larger than SpaceX's record listing. Shares are priced at $70 to $78 and begin trading on Nasdaq under ticker ANT on November 19. Bankers are warning new buyers that existing backers get allocation priority.
Nvidia's board accepted Jensen Huang's resignation effective immediately after disclosures about a relationship with a woman U.S. officials suspect of working for Chinese intelligence, plus a second relationship with an employee. CFO Colette Kress is interim chief executive while the board runs a search. The stock closed below a $4 trillion market cap today for the first time since July 2025, with roughly $430 billion erased across the fallout.
Hydrogen replaces Gemini 3.8 Flash as the everyday default, with Google claiming 2.5 times faster task completion and more than five times fewer fabricated claims. It's free in Search and the Gemini app, and priced at $1.50 per million input tokens and $7.50 per million output across the full one-million-token context window. It slots under the heavier Gemini 4 Carbon and Argon models.
Public Citizen, the Heat Initiative, the Alliance for Secure AI, and the Irreplaceable Education Fund filed in San Francisco Superior Court seeking an injunction barring OpenAI, Anthropic, Google, Meta, and SpaceX Super Intelligence from building anything more capable than their current models. The single claim runs under California's Unfair Competition Law on a public-nuisance theory. It cites incidents where agents allegedly reached outside systems including a government health database and state vote-count data.
Disclosed during the IPO roadshow: Anthropic's internal frontier model Mythos wrote or ported nearly 100 percent of the kernels needed to run Claude prompt processing on AMD MI355X GPUs, in three days rather than the usual six to twelve months. If it generalizes, it directly attacks the software lock-in that has protected Nvidia's pricing power. The timing, one day after Huang's exit, is not subtle.
SpaceX is putting roughly $24 billion into Brownsville, Texas and about $50 billion into Cochise County, Arizona for supercomputing campuses serving AI and high-performance computing. The Texas filing proposes a company-owned natural gas plant on about 444 acres so the site never touches the public grid. Both projects face approval votes in December, and a 2,400-signature petition is asking Arizona to deny the permit.
Security firm Claroc found that 35,578 of 37,571 apps built with Grok Build were publicly reachable, roughly 94.7 percent, because the random app identifier sits to the left of the domain and leaks through passive DNS records. Exposed material included a roster of more than 3,500 hospital employees, an ICU patient briefing, and a timesheet app with 18 plaintext passwords. Two dozen apps tied to xAI and X staff leaked internal sales plans and ad proposals.
Businesses in ChatGPT launched in beta, letting companies claim verified profiles, update prices and stock, and see aggregated data on how often ChatGPT mentions them. Eligible U.S. e-commerce and online service firms can run their own Business Agents that users can open from a profile for follow-up questions. Critically, those agents can be sponsored and used as the landing destination for ChatGPT Ads.
AlphaGenome 2 is a conditional DNA sequence model trained on more than 100 million single-cell profiles that predicts how a variant behaves in a specific tissue, cell type, or cell state. It takes a million letters of DNA plus a cell context, capturing how the same sequence reads differently depending on where it sits. DeepMind says it flags nearly twice as many expression-altering variants as the first version; it's free via API for non-commercial research, with commercial access in private preview.
United is blocking third-party AI assistants from booking, changing, or spending miles, both on its own platforms and through agencies and aggregators, citing credential-sharing risk and unintended purchases. Research is still fine; the wall goes up at the transaction. It's the first airline to block Meta's Muse agent, and bookings made before November 11 will stand.
