Anthropic, the maker of the Claude chatbot, said Sunday it will commit 5% of its shares to a newly created "Anthropic Public Benefit Fund," a move that could eventually funnel tens of billions of dollars toward nonprofits and people affected by the economic upheaval of artificial intelligence — just as the company gears up for an initial public offering that could be one of the largest in Wall Street history.

The pledge, detailed in a post on Anthropic's website, is subject to shareholder approval. The company said its board could tack on an additional 2% for every 50% jump in Anthropic's valuation above a baseline still to be disclosed, meaning the fund's slice of the company could keep growing alongside its fortunes.

Combined with earlier promises — cofounders' pledge to give away 80% of their personal equity, plus a separate employee donation-matching program — Anthropic says roughly 20% of its total shares are now earmarked for "the public good," with more potentially on the way.

A war chest for AI's fallout

Anthropic has not put a dollar figure on the new fund, but the math is eye-popping. The company's most recent fundraising round, a $65 billion Series H in May that was reported by law firm Orrick and other outlets, valued Anthropic at roughly $965 billion. At that valuation, a 5% stake alone would be worth close to $48 billion — though Anthropic's actual worth, and the fund's, will move with whatever price the company fetches when it finally goes public.

The fund will be structured as a company-owned LLC rather than a traditional 501(c)3 charity, which Anthropic says gives it more flexibility — including the ability to make direct cash distributions to specific groups of people, not just grants to nonprofits. The company was explicit that none of the money can be used for lobbying or political giving, in the U.S. or abroad.

Anthropic pointed to the labor market as an early test case. The company has already committed $200 million to research on how AI reshapes jobs, but said in its announcement that "bolder approaches may be needed" given how fast the technology is moving — language that leaves the door open to the fund eventually cutting checks directly to workers whose jobs are automated away.

"A new entity designed so that its stake grows as the company grows" — Anthropic, describing the fund's structure.

The timing is notable. The Associated Press reported that Anthropic is preparing for a massive IPO that could come as soon as November, which would make Sunday's pledge one of the last major announcements the closely held company makes before its finances and governance come under the kind of scrutiny public markets demand. Anthropic was founded in 2021 by CEO Dario Amodei, his sister Daniela Amodei and several other former OpenAI researchers, and has structured itself from the start as a public benefit corporation — a legal form meant to balance shareholder returns against a stated social mission.

Whether the new fund changes how skeptics view that balancing act remains to be seen. Unite.AI's writeup of the announcement noted that Anthropic has committed to publishing a "public benefit report" tracking its progress — a disclosure that, once the company is public, investors and critics alike will be able to hold it to.