Wednesday packed a markets story, a science story, a travel story, and a legal story into one day. Here’s the tech news November 12, 2026 has in store.
Key Takeaways
- Anthropic’s IPO is multiple times oversubscribed, with up to $90.4 billion on the table — potentially the largest capital raise ever.
- DeepMind’s AlphaGenome 2 predicts how DNA variants act in specific cell types, finding nearly twice as many gene-changing variants as its predecessor.
- Runway open-sourced Praxis-1, a 27-billion-parameter robot model anyone can download, fine-tune, and use commercially.
- United Airlines banned third-party AI agents from booking its flights, the first US carrier to do so.
- New York’s RAISE Act opened registration: big AI developers must sign up by January 1, 2027, or face fines of up to $3 million.
1. Anthropic’s IPO Is Swamped — and TSMC Just Bought Into Both Labs
Investor demand for Anthropic’s debut has run to several times the shares on offer, according to Bloomberg, with big institutions driving the interest — underwriters now expect pricing above the $70-to-$78 range.
Anthropic plans to sell 825 million shares, raising up to $64.35 billion, with pricing November 18 and trading starting November 19 on Nasdaq under the ticker ANT, according to the company’s prospectus. Add a $26 billion private placement from seven chip and cloud suppliers — Google and Nvidia at roughly $10 billion each — and the total could reach $90.4 billion, topping SpaceX’s record. TSMC’s board also approved $500 million each for Anthropic and OpenAI, per the Wall Street Journal via Reuters.
2. DeepMind’s AlphaGenome 2 Reads DNA One Cell at a Time
Google DeepMind announced AlphaGenome 2 on November 11. It predicts how a DNA variant behaves in one specific cell type at a time, instead of averaging across whole tissues, after training on more than 100 million cell profiles across 68 tissues.
Why does the cell-by-cell view matter? Most disease-linked genetic variants sit outside protein-coding genes, and their effect depends entirely on which cell they’re in. On a single-cell map of 28 immune cell types, it identified 30% of the variants believed to change gene expression, against 16% for the original model — nearly twice as many. DeepMind notes the model is research-only, not validated for clinical use, and the figures come from an unreviewed preprint. Still, for anyone tracking AI in healthcare, this is the most concrete step yet from AI that reads genomes to AI that explains them.
3. Runway Gives Away Its Robot Brain
Runway has released the weights for Praxis-1, its 27-billion-parameter “world-action” robot model, under the Apache 2.0 license — anyone can download, run, fine-tune, and even use it commercially. No license negotiation, no API gatekeeper.
Praxis-1 is a mixture-of-experts diffusion transformer that predicts video and robot actions together. On Nvidia’s RoboLab-120 benchmark of 1,200 tabletop episodes, the DROID fine-tuned version scored 46.2% success, ahead of FLUX 3 Action at 42.9% — still failing more than half the time, and the numbers are Runway’s own, with no independent replication yet. The release is unusually complete: three checkpoints, code on GitHub and Hugging Face, native LeRobot support, and serving through vLLM-Omni. Open-weight robotics is following the same path open-weight language models took — and after yesterday’s Claude Fable 5.2 news showed AI finishing real freelance work far better, the same curve is now coming for physical work.
4. United to the AI Agents: You Can Look, You Can’t Buy
United Airlines became the first major US airline to ban third-party AI agents from booking flights, revising its Contract of Carriage on November 11. New Rule 6.J.6 adds “the use of AI Agents to make, modify or service a Ticket, and/or to purchase any service or product” to the prohibited practices, and United can now cancel agent-made reservations without notice, according to Simple Flying, which notes the Wall Street Journal first reported the change.
The definition is broad: any software that autonomously finds, books, buys, or manages a United product — even if a human supervises every step. United cites data security and federal check-in rules. Two carve-outs: AI agents can still search and display United fares, as long as the customer completes the purchase, and bookings made before November 11 will still be honored. Licensed travel sites like Google Flights, Kayak, and Hopper are unaffected.
5. New York Opens the Register: The RAISE Act Gets Real
New York opened its RAISE Act registration portal on Tuesday. Large frontier AI developers — companies with more than $500 million in annual revenue — must register by January 1, 2027, when the law takes effect, according to TechCrunch.
Once registered, covered developers must publish safety frameworks, report critical safety incidents within 72 hours, and issue public model transparency reports. A new state office — DIGIT, the Office of Digital Innovation, Governance, Integrity, and Trust — will run the program, with fines up to $1 million for a first violation and $3 million for later ones. OpenAI and Anthropic are explicitly in scope — a strange week for Anthropic, which is heading for the biggest IPO in history while being told to file paperwork with Albany.
The Takeaway
AI is leaving the demo phase and colliding with the real world’s rules. Anthropic’s IPO shows the market still believes. AlphaGenome 2 and Praxis-1 show the technology spreading into science and robotics. And United and New York show the pushback — airlines and states writing down exactly where the agents stop.
Check back tomorrow for the next roundup.
FAQ
What is Anthropic’s IPO price range?
Anthropic plans to sell 825 million shares at $70 to $78 each, raising up to $64.35 billion from the public offering. Pricing is set for November 18, 2026, with trading starting November 19 on Nasdaq under the ticker ANT.
What does AlphaGenome 2 do?
AlphaGenome 2 is Google DeepMind’s AI model that predicts how DNA variants affect gene activity in specific human cell types. It was trained on more than 100 million cell profiles across 68 tissues and is available free for non-commercial research through an API.
Can an AI agent book a United Airlines flight?
No. As of November 11, 2026, United’s Contract of Carriage prohibits third-party AI agents from booking, modifying, or servicing tickets. AI agents can still search and display fares, but the customer must complete the purchase themselves.
What is New York’s RAISE Act?
The RAISE Act (Responsible AI Safety and Education Act) is New York’s AI safety law. Large frontier AI developers with more than $500 million in annual revenue must register by January 1, 2027, publish safety frameworks, report critical incidents within 72 hours, and issue transparency reports.
Sources
- Bloomberg, via Seoul Economic Daily — “Anthropic IPO Demand Swamps Order Book, May Top SpaceX Record” — https://en.sedaily.com/world/2026/11/12/anthropic-ipo-demand-swamps-order-book-may-top-spacex-record
- Anthropic — IPO Prospectus (November 10, 2026) — https://ipo.anthropic.com/documents/us/prospectus.pdf
- Google DeepMind — “AlphaGenome 2: Deciphering how cells read DNA to better understand disease” (November 11, 2026) — https://deepmind.google/blog/alphagenome-2-deciphering-how-cells-read-dna-to-better-understand-disease/
- Runway — “Praxis-1 Is Now Available” (November 11, 2026) — https://runway.com/news/company-news/praxis-1-is-now-available
- Simple Flying — “United Airlines Officially Bans AI Agents From Booking Your Flights And Checking You In” (November 12, 2026) — https://simpleflying.com/united-airlines-bans-ai-agents-booking-flights-check-in/
- TechCrunch — “OpenAI, Anthropic face new oversight as New York’s AI safety law takes effect” (November 10, 2026) — https://techcrunch.com/2026/11/10/openai-anthropic-face-new-oversight-as-new-yorks-ai-safety-law-takes-effect/
