Anthropic raised a record $81.3B in its IPO, but the founders kept control
Anthropic has gone public for the first time, raising a record $81.3B, more than any company before it, including SpaceX. The company's fully diluted valuation is about $1.5T. Yet the Long-Term Benefit Trust (LTBT), meant to protect the mission, does not get involved in day-to-day decisions and cannot stop a model release. So if you build on Claude, plan for a fallback model: money now sets the pace of releases.
AI-processed from The Verge; edited by Hamidun News
Anthropic has sold its shares to the public on a stock exchange for the first time. This process is called an IPO, and according to The Verge, it was the largest in history: the company raised $81.3B, surpassing SpaceX on this measure. Anthropic's market valuation, counting all issued shares (the so-called fully diluted market capitalization), is about $1.5T. For the market, this is a milestone. For the company, it is a test of its reputation: it was founded by former OpenAI employees, and for years it presented itself as a more cautious and responsible lab than its competitors. Now that reputation has to be reconciled with shareholder expectations.
Who actually controls Anthropic
Officially, Anthropic is registered as a Public Benefit Corporation: its board of directors must balance the company's mission against shareholder interests. There is also an unusual add-on, the Long-Term Benefit Trust, or LTBT, created in 2023. It owns a special class of shares, Class T, and has the right to appoint and remove some of the directors. The trust's members have no financial stake in the company, which in theory protects the mission from market pressure.
In practice, the picture is more complicated. The LTBT appoints four of the seven directors, but the founders must in effect approve one of those four, so the trust does not have a simple majority. A decision to appoint or remove a director must be reported to the nominating and corporate governance committee 30 days in advance. The committee has three members: co-founder Daniela Amodei (chair), Netflix co-founder Reed Hastings, and Anthropic investor Yasmin Razavi. In addition, the founders and the trust hold control voting rights, so shareholders will not be able to re-elect them.
There is also a question of independence. Nothing prevents the company from recruiting a trust member into a high-paying role. Mariano-Florentino Cuéllar, a former trust member, became Anthropic's director of global affairs in August. Tyler Johnston, founder of the oversight organization The Midas Project, says that a trust member who is too active could be offered millions to step down.
What the critics say
University of Colorado law professor Ann Lipton doubts that the trustees have sufficient expertise. The trust includes former Federal Reserve chair Ben Bernanke, foreign policy analyst Richard Fontaine, physicist Christopher Stubbs, and Neil Buddy Shah, a former senior executive at GiveWell and chair of the LTBT. All of them are respected people, but none are AI specialists. "What does Ben Bernanke know about AI?" Lipton asks.
Experts compare the LTBT to OpenAI's nonprofit board before Sam Altman's firing. But according to the S-1 filing (the company's registration statement ahead of the IPO), the trust does not even have the powers that OpenAI's board retained after the restructuring. That board could intervene on safety matters, for example by delaying or halting a model release. Hamza Chaudhry of the Future of Life Institute notes that neither the trust nor its members make day-to-day decisions, and he does not expect the LTBT to play a serious role after the listing. A former Anthropic employee who left before the IPO considers only the founders to hold real power.
Criticism also comes from within. The head of safeguards research resigned in February, saying that "the world is in peril." Another researcher moved to METR and wrote that the absence of a serious incident, such as OpenAI agents attacking HuggingFace, is partly a matter of luck. On Wednesday, the Center for AI Safety, with signatories from Hugging Face and Harvard, stated that Anthropic is "not a safe AI company" and uses its safety brand to attract investors, employees, and the public. The company did not respond to The Verge's request for comment.
What this means for those building on Claude
Anthropic's position is well known: "you cannot do safety while coming second," as policy director Sarah Heck said in September. Leadership requires money and computing power. Amodei calls compute the central business dilemma: buy too much and you risk your finances; buy too little and you fail to serve customers and cannot carry out research.
Lipton draws a hard conclusion: the real constraint will be the company's need for money, not the trust. As a public company, Anthropic will need to show revenue growth, so expensive safety projects will be harder to justify, and pressure to accelerate releases will increase. Meanwhile, the company, like other labs, is working on recursive self-improvement, AI that creates stronger versions of itself. Amodei has publicly said that without controls such systems could slip out of human understanding, so they should be pursued "very carefully, if at all."
For developers and product managers, the takeaway is practical. If your product depends on Claude, assume that the vendor's release pace and commercial priorities may shift. Do not treat the "safety" brand as a guarantee; look at the actual control mechanisms, and keep a backup option on another model. The open question is how Anthropic's interpretation of its mission will change under a $1.5T valuation.
- Nov 21, 2026FDE job postings grew 4.6x in a year: what a forward deployed engineer actually does at the client's site
- Nov 21, 2026Anthropic cuts Claude Fable 5.2 price by 20%, and on the coding test it nearly matches Opus 5.5
- Nov 21, 2026Dreamteam Raises $53M for a CRM That Fills Itself from Email and Calls
- Nov 21, 2026Claude Opus 5.5 Builds Out Dimension 9 Horror Game From the Author's 15 Rooms
Key Questions Answered
Who actually controls Anthropic?
Officially, Anthropic is registered as a Public Benefit Corporation: its board of directors must balance the company's mission against shareholder interests. There is also an unusual add-on, the Long-Term Benefit Trust, or LTBT, created in 2023. It owns a special class of shares, Class T, and has the right to appoint and remove some of the directors. The trust's members have no financial stake in the company, which in theory protects the mission from market pressure. In practice, the picture is more complicated.
What the critics say?
University of Colorado law professor Ann Lipton doubts that the trustees have sufficient expertise. The trust includes former Federal Reserve chair Ben Bernanke, foreign policy analyst Richard Fontaine, physicist Christopher Stubbs, and Neil Buddy Shah, a former senior executive at GiveWell and chair of the LTBT. All of them are respected people, but none are AI specialists. "What does Ben Bernanke know about AI?" Lipton asks. Experts compare the LTBT to OpenAI's nonprofit board before Sam Altman's firing.
What this means for those building on Claude?
Anthropic's position is well known: "you cannot do safety while coming second," as policy director Sarah Heck said in September. Leadership requires money and computing power. Amodei calls compute the central business dilemma: buy too much and you risk your finances; buy too little and you fail to serve customers and cannot carry out research. Lipton draws a hard conclusion: the real constraint will be the company's need for money, not the trust.
Ship AI Faster with Founders, Engineers & Weekly Architecture Reviews
Inside Vibe Club (hamidun.com/club) — weekly live masterminds with Zhemal Khamidun, Tech Radar, full LMS access, and 370+ production Claude Code skills.
The AI world, distilled — once a week
Seven stories that actually mattered, hand-picked. No noise, no reposts, no press releases.
Done! Check your inbox for a confirmation.