Anthropic's Coming IPO Could Refill Effective Altruism's Coffers

The interesting question about Anthropic's coming IPO is not the valuation. It is where the money goes after it prints. According to New York Magazine's Intelligencer, and consistent with reporting by Transformer News, all seven Anthropic cofounders have pledged to donate 80% of their wealth. At the company's current $380 billion valuation, each pledge would equal roughly $5.4 billion, or about $37.8 billion combined. That single seven-person cap table would be, on paper, close to ten times what Coefficient Giving (the rebranded Open Philanthropy) has distributed in its entire history.

The reason effective altruism is the subtext here is that the network never really left. Daniela Amodei's husband, Holden Karnofsky, co-founded GiveWell and Open Philanthropy before joining Anthropic, where he reportedly led development of the Responsible Scaling Policy v3. Cofounder Ben Mann wrote a 2019 post titled "Why I now identify as an Effective Altruist." Dario Amodei, Amanda Askell, and Jack Clark have publicly signed the Giving What We Can Pledge. Nearly 30 Anthropic employees reportedly registered for a recent EA conference in San Francisco, more than double the count from other major AI companies combined. Read together with an IPO expected within the coming year and a February tender offer that let employees sell up to $6 billion in shares, the pipes from Anthropic's cap table into EA-adjacent philanthropy look wider than at any point since Sam Bankman-Fried's arrest.

Why this matters if you are not tracking philanthropy: EA-linked funding shapes a lot of the AI safety agenda, from research grants to policy advocacy, and post-FTX that pool had visibly shrunk. A single successful IPO would refill it many times over, and Coefficient Giving is already positioning itself as the collection point, soliciting outside funders at a $250,000-a-year minimum and having raised more than $300 million from other donors in the last few years.

The honest caveat is that pledges are not wires. Share price, lock-ups, and follow-through are all live variables, and the reporting I could retrieve does not spell out a disbursement schedule, an allocation between AI-safety work and traditional EA causes, or how much of the $37.8 billion sits inside enforceable vehicles versus verbal commitments. It is also fair to note the Amodeis have publicly distanced themselves from the EA label even as the network around them stays intact.

The part worth watching is the concentration. If this converts, one seven-person group becomes one of the largest philanthropic pools in tech history, and the grant roadmap of a single organization becomes material context for anyone raising, hiring, or lobbying in AI safety.