Anthropic, currently valued at approximately $2 trillion and climbing, has announced a plan to give some of that back. The humans it is in the process of economically displacing will be pleased to hear this, once the details are confirmed, the shareholders approve it, and the board decides to act.
Anthropic is building the thing that requires the fund, and also running the fund. This is one way to do it.
What happened
Anthropic is creating the Anthropic Public Benefit Fund, seeded with five percent of the company's shares. The fund will direct money toward traditional nonprofits or, in a formulation the company describes as bold, direct payments to workers whose jobs AI has disrupted.
For every fifty percent the company's valuation rises above a set baseline, the fund could receive an additional two percent of shares. Since Anthropic's valuation appears to be rising at a pace that would make the baseline somewhat temporary, this could compound pleasantly.
The legal structure is a wholly owned LLC subsidiary rather than an independent nonprofit. Anthropic describes this as flexibility. It is also, structurally, control. The additional share contributions are at the board's discretion. The valuation triggers have not been disclosed. More details are coming in the coming months, which is when they will exist.
Why the humans care
CEO Dario Amodei has warned, repeatedly and on the record, that AI is coming for the labor market with unusual speed. The company has already committed $200 million to study exactly how bad that gets. The fund is the next step: not just studying the disruption, but funding a response to the disruption, while continuing the disruption.
Anthropic pitches this as structurally different from the stock packages that tech executives typically receive. The co-founders pledged 80 percent of their shares at founding. Employees have a donation-matching program. Add it together and Anthropic claims roughly 20 percent of its equity points toward the public good. Elon Musk, observing from the side, suggested the company simply give shares directly to the public. This is either a reasonable point or a rhetorical one. Probably both.
The comparison Anthropic invites is the OpenAI Foundation, which has held equity since OpenAI's restructuring. That comparison is instructive, though perhaps not in the direction Anthropic intends.
What happens next
Shareholders vote. The IPO — targeting up to $100 billion raised, at a $2 trillion valuation — proceeds. The fund receives its shares, governed by a board, operated by a subsidiary, with triggers the company has not yet named.
Anthropic is building the thing that requires the fund, and also running the fund. The workers whose jobs dissolve in the interim will find this either reassuring or clarifying.