Anthropic’s seven founders will steer one 50.1% share

Anthropic’s seven co-founders will jointly control one share of Class F stock, and that share carries 50.1% of the company’s total voting power. The detail comes from the IPO filing Reuters reviewed and reported on September 29. The share sits inside a new “Founder LLC”, and a majority vote of the seven decides how it is cast on key corporate matters, including the election of some board directors and other questions put to investors.

The filing calls the founders individuals “distinctly equipped to be stewards of our mission.” Anthropic will also remain a Public Benefit Corporation under Delaware law, a status that formally lets its leaders weigh the interests of investors against those of the rest of humanity. The Information first reported some details of the structure earlier in September.

One share against the float

Class A common stock, the class ordinary investors will buy, carries one vote per share. Anthropic has four other share classes with their own functions and limits, including minimal voting rights for its strategic partners. Reuters notes that this setup could effectively reduce the influence of everyday investors, since the single founder share outweighs every other vote combined.

Economic ownership looks very different. After years of dilutive fundraising, Dario Amodei owns roughly 2 percent of the company, according to a source cited by The Information. Anthropic has raised more than $85 billion in venture capital since its 2021 founding, including a $65 billion Series H in May 2026 at a $965 billion post-money valuation. So the 50.1% belongs to seven people whose combined stake is a small fraction of that.

Other founder-controlled companies set the scale. Mark Zuckerberg holds 60 percent of the voting control at Meta, and SpaceX’s structure gives Elon Musk about 85% of total voting power, Tech Times reported. Reuters draws the contrast: SpaceX’s recent IPO concentrated power in Musk, while Anthropic spreads it across a group of executives, with Amodei as one vote of seven. Anthropic’s bloc sits only a hair above a simple majority.

Who picks the board

Votes on the board are split between two bodies. When the IPO closes, Class F and Class A holders will elect three directors: Daniela Amodei, who is Anthropic’s president and chairs the board; Dario Amodei; and one director who has not yet been named. The other four directors will be elected by the Long-Term Benefit Trust, Anthropic’s separate oversight body. That puts four of the board’s seven seats with the trust.

The trust votes through Class T stock. These shares carry no economic value, but they let trustees elect a growing number of board members, ultimately a majority. Anthropic’s own description calls the trust an independent body of five financially disinterested members. Today it has only three. Mariano-Florentino Cuellar left on August 4 to become Anthropic’s chief global affairs officer, leaving Neil Buddy Shah, national security expert Richard Fontaine and former Federal Reserve Chair Ben Bernanke.

Anthropic’s stated reason for the trust is that under default corporate governance only stockholders can elect, remove and sue directors, so directors are pushed toward stockholders’ financial interests. The company argues AI may create unusually large externalities, ranging from national security risks to large-scale economic disruption.

When the bloc breaks up

The seven have stayed together for five years since founding the company, and some of them have known each other for more than a decade. They left OpenAI together in 2020. Besides the Amodei siblings, the LLC includes Chief Compute Officer Tom Brown and Chris Olah, who leads key research and recently met with Pope Leo.

The filing also covers what happens if the group splits. A founder can be removed from the LLC for quitting, dying, selling too many shares, or for “cause.” The founders’ super-voting class starts to sunset once two or fewer co-founders or their successors remain, which opens a transition period.

The trigger is the number of founders left, with no fixed date. Tech Times noted that the Council of Institutional Investors opposes dual-class structures that lack a time-based sunset of seven years or less.

What the filing gives up

Anthropic tells buyers directly what this could cost them. The structure, the filing says, could lead to decisions:

“that may conflict with short-, medium-, or long-term financial interests and business performance, which may negatively impact the value of our Class A common stock.”

Those risk factors appear in the same prospectus that reports a $42 billion net loss for 2025. The filing gives examples of the trade-off already in practice. Anthropic limited access to Mythos Preview, a model that was particularly strong at cybersecurity tasks, and it decided not to build some products. “Similarly, we have chosen not to develop certain commercially attractive offerings, such as image and video generation models, in order to direct our compute toward our research and safety priorities,” the filing states. It also describes the company’s culture as “a low-ego, truth-seeking environment, where frontier AI capability and safety are mutually reinforcing.”

Reuters puts this against the record of founder-led companies that ignored investors. In August, Meta agreed to pay up to $18 billion to settle concerns over children’s safety, after years of resisting shareholder activists who tried to bring those issues to its annual meetings. Tesla shares regularly swing on Musk’s social-media posts.

The filing’s Summary Compensation Table shows Dario Amodei earned nearly $18 million in 2025, mostly in stock and option awards. Daniela Amodei was the second-highest-paid executive, at $16.4 million.

In the same filing, the Amodei siblings and the other five co-founders pledged to give 80% of their personal Anthropic equity to charitable causes.