Anthropic’s prospective public-market investors must reckon with an external group of trustees that control the majority of the AI company’s board, as its planned blockbuster initial public offering forces close scrutiny of its experimental governance structure.
The company’s Long-Term Benefit Trust (LTBT) is a small group of advisers created to safeguard the lab’s mission of developing AI for the long-term benefit of humanity, even as commercial pressures intensify.
The trust holds no equity in Anthropic, but has significant influence, with the San Francisco-based company planning to preserve its role after a stock market debut that could value the Claude maker at as much as $2 trillion.
The LTBT, which is the controlling governance mechanism of the Anthropic public benefit corporation, has the right to appoint or dismiss the majority of the company’s board. It has selected four of Anthropic’s seven directors, including Netflix’s co-founder Reed Hastings and Vas Narasimhan, chief executive of Novartis.
Currently, the LTBT has three members out of a potential maximum of five, chaired by Neil Buddy Shah, chief executive of the Clinton Health Access Initiative, alongside former Federal Reserve chair Ben Bernanke and Richard Fontaine, CEO of the Center for a New American Security. Mariano-Florentino “Tino” Cuéllar, a former California Supreme Court justice, left the trust after a few months to work as Anthropic’s chief global affairs officer.
Anthropic is attempting to establish a template for future AI governance through the trust, said multiple people close to its plans.
They said the ambition for the LTBT is to provide an industry blueprint, similar to the GAAP accounting principles, which initially emerged as voluntary standards proposed by private companies. Bernanke’s appointment in July signals this intention to build a more institutional body, one person added.
The trustees are required to receive advance notice of major company actions, including the launch of new AI models. They meet weekly among themselves, while also meeting Anthropic’s leadership as frequently as every other week, according to people close to the matter.
They also attend regular company board meetings and hold discussions with Anthropic’s founders on significant issues. Those have included the company’s Mythos cybersecurity model, where trustees encouraged a limited rollout through the Glasswing Project, as well as its dispute with the US government over automated weapons, said a person close to the talks.
Yet despite that access and authority, the trust has operated largely in an advisory capacity. It has not attempted to draw red lines or force a significant trade-off between profit and purpose, said another person familiar with its workings—meaning the structure has yet to face the kind of conflict that would show whether it can constrain Anthropic’s leadership when commercial and societal goals diverge.
That structure creates a “built-in conflict,” according to Jesse Fried, professor at Harvard Law School and a corporate governance expert.
The company “raises funds from profit-seeking investors, then lets self-appointed individuals decide how much profit to sacrifice for the firm’s mission—ensuring its AI benefits humanity”, he wrote in a recent paper highlighting the risks of the governance models at Anthropic and OpenAI. “A deep and potentially unmanageable tension is thus hard-wired into the firms’ corporate DNA.”
Experts warn that while Anthropic’s trust is structured thoughtfully, it remains an unproven model that will need to be stress-tested as the lossmaking company moves toward becoming a sustainable business over the long term.
“It’s nearly impossible to perfectly contract for all possible circumstances that could arise when managing competing interests within a firm,” said Elizabeth Pollman, a law professor at the University of Pennsylvania and an expert in corporate governance.
She pointed to the intense competition in AI at the geopolitical and corporate levels as factors that make the balancing act even more difficult. “Will this governance structure work in the way intended, serving dual or more interests over time? That’s the real challenge,” she said.
In recent years, there has been a broader trend in Silicon Valley of tech companies diverging from conventional governance norms, making it harder for shareholders and the public to hold them to account.
Anthropic and OpenAI, the two major AI labs, have both designed unusual company structures, with self-appointed “guardians” of their mission rather than just traditional board directors with fiduciary duties.
OpenAI, which was founded as a nonprofit, became a cautionary tale in November 2023, when its board tried to fire its chief executive Sam Altman but lost the confidence of the company’s investors and employees. The incident led to most of the board members being replaced and a broader restructuring of the startup.
Anthropic’s trust is seen as less risky than OpenAI’s governance because it has a built-in “kill switch” that allows the trustees to be fired with the support of 85 percent of the shareholders’ voting power, a supermajority that could change when the company goes public, according to a person with knowledge of its structure.
A person close to Anthropic said the private investors who have backed the company across multiple rounds did so with full understanding of its governance structure, and several specifically cited its emphasis on safety as part of their investment thesis.
However, early investors also assumed the company would have to become a commercial juggernaut in order to fulfill its core mission, said one venture capitalist who has backed Anthropic.
“There was a judgment made by investors that capitalism would win in the end. Whatever you say, if you need a lot of money for compute and to compete [for the best model], investors assume there will be a business,” they said.
The move into public markets will expose that structure to a broader and potentially less-forgiving investor base as Anthropic faces mounting pressure to earn a profit.
“Under the firms’ current structures, these directors … may have little skin in the game and can or must ignore profit in decision-making. OpenAI has already had a … debacle. Anthropic, with a less risky structure, hasn’t,” wrote Harvard’s Fried in his July paper. “Investors should scrutinize both companies’ arrangements, which may still change before their IPOs, and price shares accordingly.”
