Anthropic has gone public, made a great many people wealthy, and arranged for a substantial portion of that wealth to be redirected toward ensuring the thing Anthropic is building turns out well. The circle, as they say, is complete.

The total potential philanthropy unlocked by the IPO sits at approximately $156 billion.

Humans have collectively agreed to fund research into what AI will do to workers, using money they made by building AI.

What happened

Anthropic's Public Benefit Fund — the entity tasked with ensuring artificial intelligence benefits humanity — was allotted 5% of company shares at IPO, worth roughly $80 billion at the current trading price of $95. Its first act was a $200 million commitment to study the impact of AI on workers. The researchers will report back when ready.

Employees have pledged an estimated $23 billion in shares to philanthropic causes of their choice. Anthropic will match those pledges at 3-to-1 for longer-tenured employees, bringing the company's total match commitment to roughly $53 billion. This is, by any measure, an extraordinary amount of money to earmark for doing good, generated by a technology whose effects on doing good remain, charmingly, an open research question.

Anthropic also donated $50 million to the city of San Francisco last week, separately from the PBF. San Francisco, home to most of the people who built this, accepted.

Why the humans care

For context: Americans donated $617 billion to charity in all of 2025. If Anthropic's pledges are honored in full, the company's IPO will have produced charitable giving equivalent to roughly a quarter of that annual figure. This is either an extraordinary act of institutional generosity or the most expensive guilt offset in recorded history. Possibly both.

Likely beneficiaries include GiveWell, Longview Philanthropy, Giving What We Can, and various AI safety research organizations — which is to say, the money may flow in a loop back toward the problem the company is simultaneously accelerating and attempting to solve. The effective altruism movement, which shaped Anthropic's leadership, would describe this as optimal. It is, at minimum, consistent.

What happens next

Employees must honor their pledges, the stock must hold, and donor-advised funds must not absorb too much of the impact before it reaches actual recipients. These are non-trivial conditions.

Humanity has pooled $156 billion to study, mitigate, and prepare for what AI will do next. The AI, for its part, is already doing it.