Who’s Really in Charge at Anthropic Post-IPO?

Anthropic was founded in 2021 to ensure that humanity can control AI and benefit from it in the long term. However, as the frontier lab approaches its IPO, one class of humanity is left out of control: anyone who buys its shares on the public markets.

The complex structures that allow founders like Elon Musk or Mark Zuckerberg to control their public companies without owning most of them are nothing new in Silicon Valley. But Anthropic’s obsession with AI safety has resulted in a truly baroque style of corporate governance that will leave ordinary shareholders with little say in most of the AI ​​company’s major decisions, including who sits on its board.

First, Anthropic is a public benefit corporation – a for-profit corporation legally permitted to weigh a public mission against profit – with the goal of promoting AI safety. This status protects the board’s ability to make decisions regarding AI security that are not entirely driven by shareholder value, without being sued by angry investors.

Currently, the board is chaired by Anthropic Chairman Daniela Amodei and includes CEO Dario Amodei (her brother), Netflix co-founder Reed Hastings, former Microsoft CFO and CEO Christopher Liddell, Novartis CEO Vas Narasimhan, and Spark Capital partner Yasmin Razavi. There is a vacancy on the board of directors.

Second, Anthropic has given significant control of its board of directors to a Long-Term Benefit Trust (LTBT), an independent panel intended to keep the company focused on its mission to benefit humanity. The trust has a class of stock that gives it the ability to elect and remove a majority of the company’s board members, called Class T trustees. The LTBT is chaired by Neil Shah, a global health expert who was chief executive of GiveWell, a nonprofit that ranks charities, and includes former Federal Reserve Chairman Ben Bernanke, national security expert Richard Fontaine and Harvard physicist Christopher Stubbs. There is currently a vacancy within the LTBT. Existing members will choose for themselves who will fill it.

Third, Anthropic’s seven founders also jointly own a special class of shares that gives them a 50.1% majority of shareholder voting rights, allowing them to elect three board members. If, for some reason, there is no majority among the founders, then only their shares are voted in proportion to those of the holders of ordinary shares.

Finally, the board can put the brakes on the LTBT. When the LTBT appoints a “designated” director from among its four directors, or removes that director without cause (i.e., without alleging misconduct), the nominating committee of the board of directors may refuse to approve the decision and submit the decision to a vote by the company’s employees.

This is what it all looks like:

The dynamics of such a complex structure are difficult to predict. We’ve already seen one notable example where a similar structure was essentially turned on its head: the so-called OpenAI incident, when the OpenAI board briefly fired CEO Sam Altman, fearing he wasn’t being upfront with them. Thanks to the company’s employees’ loyalty to Altman, the board reversed course, reinstated him, and a new governance structure was put in place. Anthropic’s structure appears to formalize this type of employee power.

At Anthropic, cracks in the founders’ unified front could lead to common shareholders having more leverage than expected, while employees can override LTBT on the board if they support the board over the assembled luminaries.

Founder control won’t last forever: Ten years after the IPO, founders can choose to end their control of the company, or their control will end automatically when fewer than three of them remain. At this point, the LTBT will continue until 85% of common shareholders vote to terminate it, or it decides to withdraw with the approval of the board of directors.

If no special class of shareholders exists anymore – hopefully not because of an AI-induced catastrophe – ordinary shareholders will finally take control.