In 2021, a group of employees from OpenAI left to start Anthropic, a company named to reflect its focus on humanity. This month, Anthropic is expected to make one of the largest initial public offerings in history. The IPO will test whether the company can balance its mission of safely developing AI with the pressures of being a publicly traded company. Anthropic has a unique ownership structure designed to help it stick to its original goals.
Anthropic is registered as a public benefit corporation, or PBC, which requires balancing shareholder interests with its stated mission. This structure is different from traditional companies that focus primarily on maximizing shareholder value. According to Gil Luria, head of technology research at D.A. Davidson, Anthropic's governance structure is very unique and upfront about its goals. The company's "tripartite goals" as a PBC include the pecuniary interest of stockholders, the best interest of those affected by its conduct, and the Company Purpose.
Anthropic's seven founders collectively have 50.1% of the voting rights, allowing them to control the company's operations. This is similar to a trend in Big Tech companies where founders retain majority control. In addition, Anthropic established the Long-Term Benefit Trust (LTBT) in 2023, composed of five subject-matter experts with no financial stake in the company. The LTBT has special non-financial shares with the authority to appoint four directors to the seven-person company board.
The LTBT members include Neil Buddy Shah, CEO at the Clinton Health Access Initiative, and Ben Bernanke, a former chair of the Federal Reserve. There is one vacancy in the LTBT that has yet to be filled. According to Anthropic's filing with the SEC, the LTBT may have interests that differ from shareholders and may exercise its power in a way that disagrees with or is adverse to shareholder interests.
When a company goes public, there can be pressure to grow profits and deliver strong results for shareholders. However, Anthropic's CEO, Dario Amodei, has talked about the risk of AI causing harm to humanity if placed in the wrong hands. The company's IPO filing includes over 90 pages of risk disclosures, varying from AI misuse by bad actors to existential risk to humanity.
The ownership structure of Anthropic won't matter as much if profits are strong and it follows its mission of developing safe AI. However, it will be stress-tested if earnings aren't as expected, safety issues arise, or disagreements fester between the board, founders, or trustees. According to Jay Ritter, professor emeritus at the University of Florida, public shareholders don't have power in this structure.
The implications of Anthropic's corporate governance structure will take time to become apparent. According to Jay Goldberg, senior analyst at Seaport Research Partners, this structure gives the LTBT and the founders effective permanent control of the company. The debut of Anthropic on Wall Street will be a guinea pig for AI frontier labs and how they are received by investors.
Key points
- Anthropic's unique ownership structure is designed to balance profit growth with its mission of safely developing AI.
- The company is registered as a public benefit corporation, requiring it to balance shareholder interests with its stated mission.
- The Long-Term Benefit Trust has special non-financial shares with the authority to appoint four directors to the seven-person company board.
