Anthropic promises 20% of its shares to charity, but details are vague

Anthropic, the AI company, says roughly 20 percent of its shares will go to the public good. But it has not said how much will reach charities, when the money moves, or who decides where it goes. The stock sale, which could happen before Thanksgiving, may value the company at $1.5 trillion to $2 trillion. That makes the missing details a big deal.

What Anthropic promised

There are two main pieces. First, 5 percent of shares would go into an LLC called the Anthropic Public Benefit Fund, subject to shareholder approval. Second, the seven co-founders pledged to donate 80 percent of their shares to philanthropic causes. Add a matching program for employees who give to charity, and Anthropic says the total reaches about 20 percent. The securities filing notes that the 80 percent pledge is not binding.

The size of the money

At the expected valuation, 20 percent would be $300 billion to $400 billion. The public benefit fund alone could be worth $75 billion to $100 billion. Rival OpenAI is controlled by a nonprofit that holds a 26 percent stake, which could be worth $200 billion or more. Together, the two could give $500 billion to $600 billion. That is two and a half to three times what the Gates Foundation expects to spend before it closes in 2045.

What is still unclear

• How much of the fund will actually go to charity.
• When the shares will be transferred, and how fast they can be sold.
• Who will make the decisions about how the money is spent.

For now, the fund will be run by three co-founders and Neil Buddy Shah, a global-health leader. A separate governance setup will come later. The money could fund grants to nonprofits and help people hurt by job losses, likely through direct payments. The co-founders will also hold special stock carrying 50.1 percent of the votes.

Why some experts are wary

Steven Dean, a Boston University law professor, said the company is vague on purpose. He said that suggests people should not have high expectations. Phil Buchanan of the Center for Effective Philanthropy worries the money will chase new, unproven ideas instead of backing groups that already work. He also called the case for using an LLC weak, since it is shielded from public disclosure of its grants. Other leaders are excited about what the money could do.

What happens next

An Anthropic spokesperson declined to comment but promised greater clarity in the coming weeks and months. Experts expect the money to target two risks tied to AI itself: the threat to humanity and the potential for mass layoffs. Until the company puts specifics in writing, the pledge remains a promise, not a plan.