Anthropic was established in 2021 with a stated mission of ensuring that artificial intelligence remains safe and beneficial to humanity over the long run. Yet as the company moves closer to listing its shares on public markets, a detailed look at its corporate structure reveals that ordinary investors who buy those shares will have remarkably little influence over how the company is actually run.

Dual-class share structures that allow founders to retain outsized voting power even after a company goes public are a well-established feature of Silicon Valley. Companies led by figures like Elon Musk and Mark Zuckerberg have long used such mechanisms to insulate founders from shareholder pressure. Anthropic, however, has taken this concept considerably further. Driven by its deep commitment to AI safety, the company has constructed what can only be described as an unusually layered and intricate governance framework - one that leaves common stockholders with very limited influence over most consequential decisions, including who sits on the board of directors.

A Public Benefit Company With a Safety-First Charter

The first layer of this structure begins with Anthropic's corporate classification. The company is incorporated as a public benefit corporation - a legal designation that permits a for-profit business to formally weigh social or public goals alongside financial returns. Anthropic's defined public purpose centers on the responsible development of AI. This classification is significant because it provides the board with legal cover to make decisions grounded in AI safety considerations, even when those decisions might not maximize shareholder value. Without this status, disgruntled investors could potentially sue the board for prioritizing mission over profit.

Currently, Anthropic's board is chaired by company president Daniela Amodei. It also includes her brother and CEO Dario Amodei, Netflix co-founder Reed Hastings, former Microsoft and General Motors CFO Christopher Liddell, Novartis chief executive Vas Narasimhan, and Spark Capital partner Yasmin Razavi. One board seat remains unfilled at this time.

The Long Term Benefit Trust

The second and arguably most distinctive element of Anthropic's governance is a body called the Long Term Benefit Trust, or LTBT. This is an independent panel specifically designed to keep the company anchored to its mission of benefiting humanity rather than drifting toward purely commercial priorities. The LTBT holds a dedicated class of stock that grants it the authority to elect and remove a majority of Anthropic's board members - a category formally designated as Class T directors.

The LTBT is chaired by Neil Shah, a global health specialist who previously served as managing director at GiveWell, the nonprofit organization known for evaluating charitable effectiveness. Other members include former Federal Reserve chair Ben Bernanke, national security analyst Richard Fontaine, and Harvard physicist Christopher Stubbs. Like the board itself, the LTBT also has one vacancy. Notably, the existing members of the trust - not shareholders or the board - will decide who fills that open seat.

Founder Control Through a Special Share Class

The third layer of control rests with Anthropic's seven co-founders. Together, they hold a special class of shares that collectively carries 50.1 percent of total shareholder voting power, giving them the ability to elect three members of the board of directors. This majority voting bloc remains unified as long as more than half of the founders agree on how to exercise it. If that majority breaks down for any reason, the founders' shares revert to being voted in proportion with common stockholders - a provision that introduces a degree of unpredictability into how much sway ordinary investors might actually have.

A Board Check on the Trust

The fourth mechanism introduces a counterbalance to the LTBT's power. When the trust nominates or removes - without cause, meaning without alleging any misconduct - what the company terms a "Specified" director from among its four board appointments, the board's own nominating committee has the authority to reject that decision. If the committee does decline to approve the move, the matter does not go to common shareholders. Instead, it is put to a vote among the company's employees.

This employee vote provision is one of the more unusual features of the entire structure and echoes a real-world precedent that drew significant attention in the AI industry. In late 2023, OpenAI's board briefly removed CEO Sam Altman, citing concerns that he had not been sufficiently transparent with them. The move backfired when the vast majority of OpenAI's employees rallied behind Altman and threatened to resign en masse. The board reversed course, reinstated Altman, and subsequently overhauled its own governance arrangements. Anthropic's structure appears to deliberately formalize that kind of employee influence, building it into the governance framework from the outset rather than leaving it as an informal pressure point.

Where Cracks Could Appear

Despite its elaborate design, the structure is not without potential fault lines. If the founding group were to fracture and fail to maintain a unified majority position, common shareholders could find themselves with more voting power than the framework appears to intend. Similarly, employees hold a meaningful card: if they side with the board rather than the LTBT in a dispute over a board seat, they can effectively overrule the assembled group of distinguished trust members.

The founders' control is also explicitly time-limited. A decade after Anthropic's initial public offering, the founders may voluntarily relinquish their special voting rights. If they choose not to, that control will end automatically once fewer than three of the original seven founders remain involved. At that point, the LTBT would continue to exercise its influence until either 85 percent of common shareholders vote to dissolve it, or the trust itself chooses to step down with board approval.

The Long Road to Ordinary Shareholder Power

Only when both the founders' special share class and the LTBT have ceased to exist will common shareholders gain meaningful control over Anthropic. The company's governance documents appear to contemplate that scenario arriving eventually through the natural passage of time and the decisions of those holding special rights - rather than through any action available to the investors who purchase shares on the open market.

The structure reflects a calculated bet by Anthropic's founders that insulating the company from conventional shareholder pressure is necessary to pursue what they regard as a long-term mission of global importance. Whether public market investors will accept those terms - and at what valuation - remains to be seen as the company advances toward its listing.