Anthropic plans to grant its chief executive, Dario Amodei, and the company’s other co-founders a class of shares with outsized voting power, according to two people familiar with the matter, a governance move designed to help the founders resist pressure from outside shareholders as the AI company edges toward a public listing.
The arrangement would be the first time Anthropic’s management has held super-voting shares. The mechanism is familiar in technology: Google’s founders, Meta’s Mark Zuckerberg and countless startups have used multi-class structures to keep control after going public. What is unusual at Anthropic, the people said, is the starting point. The founders’ equity stake is comparatively small for a company of its prominence, a legacy of the fast and complex fundraising rounds that built the firm. Anthropic has raised enormous sums from a parade of strategic investors, with Google, Amazon and Microsoft among those that took stakes in exchange for capital and computing credits. Each round diluted the founders further, leaving them with less of the company than their counterparts at almost any other major technology firm.
The company is also planning a second, less conventional layer of protection. Anthropic intends to use a special class of stock to preserve an existing group of non-shareholder trustees, who would continue to elect a majority of the board. The trustee body, established during the company’s earlier years, was designed to keep a set of safety commitments binding even as ownership changed. That structure, the people said, creates a barrier that does not depend on the founders’ own ownership: even if outside investors accumulate ordinary shares, the trustee group keeps command of board seats.
Taken together, the two mechanisms would give Anthropic one of the more fortress-like governance structures in the AI industry. The company is not commenting publicly on the plans. Lawyers who structure technology IPOs said the design has a further benefit for the founders: it signals to employees that control will not drift toward outside capital, a reassurance that matters when retention is paid in stock that cannot be sold until a listing.
The moves come as Anthropic approaches a moment of maximum bargaining power for investors. The company’s revenue has overtaken OpenAI’s, its operating profit has turned positive, and the market’s appetite for AI stocks is strong. A public offering would let early investors cash out, but it would also hand them a seat at the table. The share structure is the answer to that handover.
Governance specialists said the combination is notable for what it says about the company’s founders. Most tech founders who demand super-voting power arrive at the IPO with large personal stakes and argue that their shares reflect their outsize role. Amodei and his co-founders, including his sister Daniela Amodei, who serves as president, cannot make that argument with the same force, because the rounds that funded Anthropic’s research took a larger share of the company. The voting structure, the specialists said, is therefore less about entrenching an owner and more about insulating a mission.
The mission question is central to Anthropic’s identity. The company was founded on a charter that commits it to AI safety, and it has repeatedly said it wants a governance arrangement that keeps those commitments even if the company changes hands. The trustee mechanism, people close to the company said, is designed so that board control does not simply follow the largest check.
The plans are still taking shape, and the details could change before any filing. The company has not confirmed an IPO timeline, though its financial disclosures have made a listing look increasingly likely. Anthropic’s structure will also need to pass muster with the public markets: investors have accepted multi-class structures for years, but some index providers and governance firms have pushed back on companies that concentrate power too aggressively, and a founder-controlled AI company listing at a nine-figure valuation will draw scrutiny it has never faced as a private firm.
The timing is delicate for another reason. Anthropic’s rise has been built partly on its contrast with OpenAI, whose own governance has gone through repeated reworkings, including the brief ouster of its chief executive in 2023. By locking control with its founders and trustees before listing, Anthropic is trying to avoid the kind of drama that has defined its larger rival.
For the people who invest in Anthropic, the message is plain: buy the stock, but do not expect to run the company. For the founders, the structure is the difference between building what they set out to build and managing what their investors want. The vote, when it comes, will be theirs.
