Anthropic's complex governance structure limits investor power

A mission-first corporate model

Anthropic was established in 2021 with the specific goal of ensuring that humanity can control and benefit from artificial intelligence in the long term. As the company prepares for its initial public offering (IPO), it is adopting a corporate structure that prioritizes this mission over the interests of common shareholders. This approach is designed to prevent the company from being forced to make safety-compromising decisions just to satisfy investors.

To achieve this, Anthropic is organized as a public benefit company. This legal designation allows a for-profit corporation to weigh a public mission alongside its pursuit of profit. By having this status, the board can make decisions regarding AI safety that are not driven entirely by shareholder value without facing lawsuits from angry investors.

The current board is led by president Daniela Amodei and includes CEO Dario Amodei. Other notable members include Netflix cofounder Reed Hastings and former Microsoft and GM CFO Christopher Liddell. The board currently has one vacancy remaining as it prepares for its next phase of growth.

The Long Term Benefit Trust

A significant portion of Anthropic's control is allocated to a Long Term Benefit Trust (LTBT). This is an independent panel created to keep the company focused on its mission to benefit humanity. The trust owns a special class of stock that gives it the power to elect and remove a majority of the company's board members, known as Class T directors.

The LTBT is currently chaired by Neil Shah, a global health expert who previously managed a nonprofit that ranks charities. Other members include former Federal Reserve chair Ben Bernanke and national security expert Richard Fontaine. These individuals are intended to provide oversight that looks beyond the immediate financial horizon of the company.

The trust's influence is substantial because it can effectively choose the leadership of the company. However, the board can still put a check on the trust. If the trust tries to remove a director without cause, the board's nominating committee can decline the move and instead throw the decision to the company's employees for a vote.

Founder and employee influence

The seven founders of Anthropic also hold a special class of stock that provides a majority of them with 50.1% of the shareholder voting power. This specific allocation allows them to elect three members of the board of directors. This ensures that the original creators of the company maintain a significant seat at the table even as the company scales.

If the founders ever fail to reach a majority agreement, their shares are then voted in proportion with those of common stock holders. This creates a tiered system of power where the founders' interests are primary, followed by the trust's mission-based oversight, and finally the interests of the general public.

The structure also formalizes a type of employee power that we have seen in other AI labs. At Anthropic, if employees back the board over the trust's assembled luminaries, they can potentially overrule the trust on specific board seats. This creates a complex dynamic where the company's direction could be influenced by internal staff loyalty.

The path to public control

The founders' control is not intended to be permanent. Ten years after the company goes public, the founders can choose to end their control of the company. Alternatively, their control will end automatically if fewer than three of the original seven founders remain in the company.

Once the founders' special class of stock is no longer active, the LTBT will continue its oversight. However, the trust can only continue until 85% of common shareholders vote to end it, or if the trust decides to step down with board approval. This provides a long-term roadmap for how the company will eventually transition to a more traditional corporate structure.

If and when both the founders' special class and the trust cease to exist, common shareholders will finally gain full control. The current setup is designed to prevent this transition from happening prematurely, ensuring that the company's safety goals are protected during its most critical growth years.