Anthropic is designing a special stock class that would give founder and chief executive Dario Amodei, along with his co-founders, outsized voting power once the company lists publicly. Reuters reported the plan on Wednesday, relaying a scoop from The Information that cited two people with direct knowledge of the plans. The specific mechanics, including how many votes each share would carry, were not learned, and Anthropic could still revise or drop the structure before any listing. Anthropic did not respond to Reuters’ request for comment.
The arrangement would be a first for Anthropic’s leadership, which has never held supervoting stock. It also sits oddly against the ownership numbers: Amodei holds roughly 2 percent of the company, a stake far smaller than what founders at comparably sized tech firms usually carry into an IPO. Control, in other words, would be built almost entirely apart from equity.
Anthropic already runs on a second layer of control most public companies don’t have. The startup is a public benefit corporation, legally bound to weigh its stated mission alongside profit, and it maintains a Long-Term Benefit Trust: trustees who hold no equity but who oversee a share class empowered to seat most of the board’s directors. Stacking founder supervoting shares on top of that trust would mean two separate mechanisms, neither one answerable to ordinary shareholders, sitting above whatever public investors eventually buy.
That leaves a concrete question for anyone eyeing Anthropic’s eventual share sale: what exactly they would be buying control of. A board-electing trust and a supervoting founder class are both designed to override outside shareholders, but nothing in the reporting explains how the two would settle a disagreement between themselves. Reuters did not say what happens if the trustees’ read on Anthropic’s public benefit mission ever conflicts with what Amodei and his co-founders want.
Dual-class structures that concentrate control with founders are common in Silicon Valley. Meta’s Mark Zuckerberg controls close to 60 percent of that company’s voting power through his own super-voting shares, and Elon Musk’s SpaceX gives him outsized control through a comparable arrangement. Anthropic’s version would go a step further: two layers of control instead of one, with the older layer already immune to a shareholder vote by design.
Public benefit corporation status is the detail that makes the trust possible in the first place; an ordinary for-profit corporation would not install a non-shareholder body that picks most of the board. Anthropic is heading toward what Reuters described as a potential IPO later this year, one that could rank among the largest market debuts on record. A prospectus built on top of a trust and a founder share class is not a governance footnote for that filing.
Prospective investors should read the control provisions before the growth numbers. A structure with two shareholder-proof layers changes what an Anthropic share would actually confer: economic upside, not a vote on how the company gets run. Anyone evaluating an eventual Anthropic allocation should treat the Long-Term Benefit Trust and any new founder share class as the real seat of power, not the board summary on the prospectus’s opening pages.
Reuters, 19 August 2026, drawing on The Information’s original report about Anthropic’s supervoting plans.
