
KEY POINTS
- Anthropic's IPO raised $81.3 billion, which The Verge says makes it the most valuable offering of all time, ahead of SpaceX.
- Experts say the Long-Term Benefit Trust has limited authority and that the founders hold the real decision-making power.
- A fully diluted market cap of about $1.5 trillion adds pressure on a company that has built its image on safety.
Anthropic has gone public in an offering that raised $81.3 billion, which The Verge reports is the biggest IPO ever and larger than SpaceX's. The outlet says the debut also sets a precedent for other AI labs expected to list, and raises questions about how a company built on a safety mission will handle shareholder expectations.
Anthropic was founded by people who left senior research roles at OpenAI, and it has long presented itself as more careful than its rivals. Its executives have argued that the mission can only succeed if the company leads the AI race, which requires large amounts of money and computing power. The Verge notes that careful development can clash with the speed that such a race demands.
The company is registered as a Public Benefit Corporation, so its board must weigh its mission against shareholder returns. It also has an unusual structure. The Long-Term Benefit Trust, created in 2023, holds a special class of stock that lets its trustees choose and remove some board members. Together with the founders, the trust has controlling voting rights, so shareholders cannot vote the leadership out.
According to The Verge, the trust's power is narrower than it looks. It picks four of the seven board members, but one of those four effectively needs founder approval. Any nomination or removal must be announced 30 days ahead to a three-person governance committee made up of Daniela Amodei, Reed Hastings and Yasmin Razavi. The trustees are Ben Bernanke, Richard Fontaine, Christopher Stubbs and Neil Buddy Shah, who chairs the trust.
Outside experts were skeptical. Ann Lipton, a University of Colorado law professor, questioned whether the trustees know enough about AI to decide when to slow a product. Tyler Johnston of watchdog group The Midas Project noted that the company could hire away a trustee; former trustee Mariano-Florentino Cuéllar joined Anthropic in August as Chief Global Affairs Officer. Hamza Chaudhry of the Future of Life Institute said the trust has no say over day-to-day decisions.
A former Anthropic employee told the outlet that the founders hold the real power and that the trustees were chosen so as not to disrupt them. He also described the current pace of AI development as untenable under normal market structures. The Verge says Anthropic's Form S-1 suggests the trust lacks the powers that OpenAI's nonprofit board kept, such as halting a model release.
Pressure to raise money is already visible. Amodei said in a livestreamed talk with JPMorgan Chase CEO Jamie Dimon that computing capacity is the company's hardest business dilemma, because overspending and underspending both carry risks. Policy chief Sarah Heck said in September that safety cannot be achieved from second place.
Critics are not convinced. On Wednesday the Center for AI Safety published a statement with signatories from Hugging Face and Harvard University, among others, calling Anthropic not a safe AI company and accusing it of using safety branding to attract investors. The Verge's report does not include a response from Anthropic to that statement. The outlet says the open question is how the company's mission will be interpreted now that it faces a fully diluted market cap of about $1.5 trillion.