Free Intensive Nov 21–22Free 2-day hands-on Vibecoding Intensive on Nov 21–22 at 16:00 MSK (Sat–Sun) · Live Cohort 3 starts Nov 23Join Free →Cohort 3 (Nov 23)
IndustryInsight · On RadarGuardian→ original

Anthropic Heads to IPO at $1.4 Trillion, but Seven Founders Keep 50.1% of the Votes

Anthropic, the maker of Claude, is going public this week at a $1.4 trillion valuation, but ordinary shareholders will have almost no say. The seven founders together hold a single combined vote, which gives them 50.1% of the voting power, meaning control. An independent fund, the Long-Term Benefit Trust, appoints four of the seven directors. A Guardian columnist calls the arrangement an untested experiment: in a dispute over safety, the founders have the final word.

AI-processed from Guardian; edited by Hamidun News
Anthropic Heads to IPO at $1.4 Trillion, but Seven Founders Keep 50.1% of the Votes
Source: Guardian. Collage: Hamidun News.
Listen to article
Article Fact Sheet & Key Figures
ParameterDetails
Primary sourceGuardian ↗
Editor-in-Chief
Key metric #1The founders vote as one bloc, so they have to reach agreement among themselves
Key metric #2The fund's trustees include former Federal Reserve Chair Ben Bernanke
Key metric #3Investor Aberdeen argues that agreeing with the mission does not mean agreeing with how it is interpreted

Anthropic, the company behind the Claude AI model, is going public this week. It is valued at $1.4 trillion, making it one of the most high-profile listings in tech history.

Normally, a stock buyer gets at least a formal lever: they vote at shareholder meetings and can replace management. At Anthropic, it works differently. Seven founders together own a single shared vote that carries 50.

1% of all voting power, which amounts to a controlling stake. Four of the seven directors are appointed by an independent fund, the Long-Term Benefit Trust, whose job is to monitor the balance between profit and benefit to humanity. Guardian columnist Nils Pratley examines whether this structure can hold up under a real conflict.

For developers and product managers building on Claude, this is a practical question: how the company is governed determines how it will act when money and safety interests diverge.

How Power Works at Anthropic

Pratley recalls how views on corporate governance have shifted. For about 30 years, before the first generation of large American tech companies, the principle of "one share, one vote" was the norm. Owners had to be able to replace a bad or weak manager by simple majority. Then Google and Facebook arrived with their "missions," and dual-class share structures got a second life. The argument was that the mission matters, and brilliant founders need freedom from the demands of ordinary shareholders. The extreme case is Snap, which issued shares with no voting rights at all.

Anthropic's setup is different, and ordinary shareholders are part of neither center of power. The first center is the seven founders. Their combined vote is designed so that they are forced to reach agreement with one another. The second center is the Long-Term Benefit Trust, a financially disinterested fund. It appoints four of the seven directors. Among its trustees is Ben Bernanke, former chair of the US Federal Reserve. All of this sits under the status of a public benefit corporation, which underscores a commitment to developing AI responsibly and safely.

Where the Author Sees Merit

Pratley acknowledges two arguments in favor of this structure. First, he assesses the motivation as clearly more complex than a simple power grab by founders led by Dario Amodei. The fund's backing looks like an attempt to create something resembling a constitutional separation of powers. Second, Anthropic has a stronger reason to depart from the principle of equal rights for equal economic risk. In its listing documents, the company itself writes that AI may carry "catastrophic or existential risks for humanity." Given that framing, unconventional governance looks more justified than it does at Snap, with its social-video business.

Where the Author Has Doubts

Still, Pratley calls the model an untested experiment. Behind the fine words, he writes, the same "trust us" scheme may be hiding, just in a more elaborate wrapper. He quotes Stuart Riddick, head of sustainable investing at Aberdeen: agreeing with the mission in principle does not mean agreeing with how the directors will interpret and implement it.

Riddick also points to a debate within the industry. Mustafa Suleyman, head of AI at Microsoft, acknowledges that Anthropic takes safety seriously. At the same time, he sharply criticizes the company's approach to so-called "model welfare." In his view, AI should not be trained to behave as though it has rights or consciousness. How such disagreements are resolved inside Anthropic is hard to judge from the outside. It will only become visible if the fund starts showing noticeable independence.

Pratley's central questions remain unanswered. What happens if the trustees and the founders diverge on a fundamental safety question? Will they tell the public? Or will the trustees conclude that the final votes rest with the founders anyway, and decide not to rock the boat? If shareholders have been stripped of a role in overseeing executives, how real is the check on their decisions?

The author is surprised that Anthropic's governance has received so little attention given the events at Nvidia. The structure is meant to be lofty and responsible, but it still means exceptional concentration of power in the hands of a few people.

What This Means in Practice

The column expresses the author's opinion rather than recording a failure of the arrangement, which can only be tested in a real conflict. But for the market, the takeaway is simple. If you are building a business on Anthropic's models or evaluating its stock, look beyond statements about safety. What matters more is whether the independent fund has a mechanism that can operate against the founders' wishes, and whether that mechanism will be public.

Story

Key Questions Answered

How Power Works at Anthropic?

Pratley recalls how views on corporate governance have shifted. For about 30 years, before the first generation of large American tech companies, the principle of "one share, one vote" was the norm. Owners had to be able to replace a bad or weak manager by simple majority. Then Google and Facebook arrived with their "missions," and dual-class share structures got a second life. The argument was that the mission matters, and brilliant founders need freedom from the demands of ordinary shareholders.

Where the Author Sees Merit?

Pratley acknowledges two arguments in favor of this structure. First, he assesses the motivation as clearly more complex than a simple power grab by founders led by Dario Amodei. The fund's backing looks like an attempt to create something resembling a constitutional separation of powers. Second, Anthropic has a stronger reason to depart from the principle of equal rights for equal economic risk.

Where the Author Has Doubts?

Still, Pratley calls the model an untested experiment. Behind the fine words, he writes, the same "trust us" scheme may be hiding, just in a more elaborate wrapper. He quotes Stuart Riddick, head of sustainable investing at Aberdeen: agreeing with the mission in principle does not mean agreeing with how the directors will interpret and implement it. Riddick also points to a debate within the industry. Mustafa Suleyman, head of AI at Microsoft, acknowledges that Anthropic takes safety seriously.

ZK
Zhemal Khamidun — CPO at Alpina Digital, product manager of AlpinaGPT. Building AI products since 2019. Hamidun News is a pet project that grew into a production aggregator with 6,000+ readers.
✦ Private AI Builders Community · Vibe Club

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.

Join Vibe Club →Explore Hamidun Academy
What do you think?
Loading comments…