Anthropic’s $2 Trillion IPO Tests Its Unusual Board Trust

Wall Street is weeks away from stress-testing one of the most unconventional corporate governance structures in Silicon Valley history. Anthropic’s impending initial public offering, potentially valuing the artificial intelligence firm at more than $2 trillion, will force public markets to price a company where outside shareholders are legally designed to hold minimal power.

The mechanism keeping retail and institutional investors at arm’s length is the Long-Term Benefit Trust. This body holds a special class of stock carrying zero economic value but grants the right to elect a majority of Anthropic’s seven-person board of directors. The appointment of Harvard astrophysicist Christopher Stubbs to fill one of two open seats on the trust, according to the company, adds another independent voice to a panel already populated by figures such as former Federal Reserve Chair Ben Bernanke, who joined in July, and Richard Fontaine, CEO of the Center for a New American Security.

Stubbs brings a background rooted in large-scale scientific collaboration rather than corporate finance. He served as the first project scientist for what is now the Vera C. Rubin Observatory in Chile and was part of the High-Z Supernova Search Team that won the 2015 Breakthrough Prize in Fundamental Physics for demonstrating the universe’s accelerating expansion. He also chaired JASON, an independent group advising the U.S. government on defense and technology, and served as Harvard University’s Dean of Science from 2018 to 2024.

His entry into the trust follows the August departure of former California Supreme Court Justice Mariano-Florentino Cuéllar, who left the oversight panel to become Anthropic’s first chief global affairs officer. Neil Buddy Shah, CEO of the Clinton Health Access Initiative and chair of the trust, cited Stubbs’ experience leading complex endeavors across academia and national laboratories. Daniela Amodei, Anthropic’s cofounder and president, stated that Stubbs possessed a track record of following data “even when it went against the prevailing wisdom of the field.”

For prospective IPO investors, the composition of the trust matters less than its structural dominance combined with management’s voting rights. Reports indicate Anthropic has planned super-voting shares for its founders. The company confidentially filed a draft registration statement with the SEC in June, following a May funding round that pegged its valuation at $965 billion.

Operating as a Public Benefit Corporation legally obligates Anthropic to weigh shareholder returns against its stated mission of responsible AI development. Critics argue this dual mandate, enforced by an unelected trust and insulated by founder super-votes, leaves public buyers with financial exposure but virtually no strategic influence over the company’s direction.

To address mounting scrutiny over this imbalance, Anthropic announced last weekend it would allocate 5% of its shares into a separate Public Benefit Fund. Unlike the trust, which governs the board, this fund is designated for philanthropic donations and direct support for individuals negatively impacted by AI automation, including displaced workers. Stubbs, the first physicist appointed to the trust, noted in a statement that he uses AI tools daily in his research and confronts the challenges they raise as an educator, adding that he recently wrote about ensuring AI-assisted research remains reproducible when models yield different answers to identical questions.