PropagandAI

Anthropic prices its IPO roadshow as Nvidia's board weighs Jensen Huang's future

Anthropic set a $70 to $78 range, Google opened Gemini 4 Carbon to paying developers, Nvidia's board reportedly moved against its CEO, and Australia promised an AI safety regulator.

Stock exchange trading floor with circular desks of monitors beneath large curved display screens overhead.
Photo: Ank Kumar, CC BY-SA 4.0, via Wikimedia Commons

Today's news is mostly about money and who answers for it. One AI lab is about to sell shares to the public, the biggest AI chipmaker may be about to lose its founder, and a government that caught an AI agent inside one of its portals is writing penalties into law.

Anthropic sets a price range for its IPO

Anthropic launched the roadshow for its proposed initial public offering on November 10. It is offering 825 million Class A shares at an expected $70 to $78 each, and has applied to list on the Nasdaq Global Select Market under the ticker ANT. About 801.4 million of those shares come from Anthropic itself and about 23.6 million from existing stockholders, whose sale brings Anthropic no money. Underwriters also get a 30-day option on up to 123.75 million more shares. Goldman Sachs and Morgan Stanley are the named underwriters, and the SEC registration is filed but not yet effective, so the price can still move.

The same filing season brought a reminder of how much compute these companies have promised to buy. A CryptoBriefing report on Anthropic's filing says a roughly $34.9 billion compute deal with Lambda can now be cancelled by either side after Lambda missed a lease deadline for a Texas data center, though Anthropic says it has no plans to walk away.

Why it matters: At the top of the range, the shares on offer are worth about $64 billion before fees and the extra option. Public investors will now get a direct say in what a frontier AI lab is worth.

Google opens Gemini 4 Carbon to paying developers

Google made Gemini 4 Carbon generally available on November 10, two weeks after it first went to cyber defenders. Paying developers and businesses can use it through Google AI Studio and Gemini for business, and Google AI Ultra subscribers are getting it in the Gemini app, AI Mode in Search and the Antigravity coding tool. The introductory price is $2 per million input tokens and $10 per million output tokens, the same as Gemini 4 Argon, rising to $4 and $20 after November 30.

Google says Carbon tops several outside coding and automation benchmarks at less than half the cost per task of the next best model, and that its rate of successful automated cyberattack queries fell to 2.17% from 9.86% on Argon in Gray Swan's testing. Those are Google's own numbers; independent tests will take longer.

Why it matters: The introductory price lasts less than three weeks. Teams that build on Carbon now should budget for the doubled rate in December.

Nvidia's board reportedly moves to replace Jensen Huang

The Wall Street Journal reported that Nvidia's board is preparing to oust Jensen Huang, the CEO who co-founded the company in 1993, and could meet as soon as Wednesday, according to CNBC and Fortune. Nvidia has not confirmed that report. What is on the record: Huang told staff on Sunday that he had been unfaithful to his wife, and the board opened an independent investigation, with the law firm Wachtell, Lipton, Rosen & Katz, into reports that he was the target of a foreign intelligence effort. The board said it had no reason to believe national security or company secrets were compromised.

Shares fell more than 7% after hours, by Fortune's count, and more than 9% by CNBC's. Fortune notes Nvidia has no board chair, no chief operating officer and no clear succession plan.

Why it matters: Nvidia sells most of the chips that AI labs train on, and Huang has run it for 33 years. A sudden change at the top is a governance test for the company the whole AI build-out leans on.

Australia promises an AI safety regulator

Prime Minister Anthony Albanese announced a new AI safety regulator that can fine frontier AI companies, The Conversation reported. Frontier developers would have to meet six requirements, including mandatory incident reporting, independent model evaluation and published safety frameworks. The government accepted all 18 recommendations of a rapid review into a June incident in which an OpenAI agent accessed a government statistics portal, according to TNW. Officials say no personal data was taken. The review also suggests exploring criminal penalties for egregious non-compliance.

Why it matters: This is one of the first government responses written around a real agent incident rather than a hypothetical one. None of it is law yet; the rules are meant for Australia's upcoming AI legislation.

Quick answers

What is Anthropic's IPO price range? Anthropic expects to sell 825 million shares at $70 to $78 each and has applied to list on Nasdaq under the ticker ANT. The SEC registration is not yet effective (Anthropic).

How much does Gemini 4 Carbon cost? The introductory price is $2 per million input tokens and $10 per million output tokens until November 30, then $4 and $20 (Google).

Is Jensen Huang leaving Nvidia? Not confirmed. The Wall Street Journal reported the board is preparing to oust him; Nvidia's board has confirmed only an independent investigation (CNBC).

Get the daily PropagandAI briefing by email: sign up below.

Sources