Public benefit fund
Anthropic’s Public Benefit Fund (PBF), comparable to OpenAI’s fund, received 5% of the company’s shares — nearly 850 million. At the current price of about $95 a share, its assets would be worth $80 billion.
If the company continues to grow, the fund’s stake could increase by another 2 percentage points each time Anthropic’s valuation rises by more than 50% relative to its IPO price of $83 a share. The threshold for the next increase is about $124.50 a share.
The fund will direct money to research and projects focused on increasing AI’s benefits for humanity. The PBF’s first commitment is $200 million to study AI’s effects on workers. Anthropic has also said it will consider donations to charitable organizations and “direct payments to specific groups of people”; TechCrunch suggests this could refer to future plans for universal basic income.
Last week, Anthropic donated $50 million to San Francisco. A company representative told TechCrunch that the money did not come from the PBF.
Employee donations
Anthropic employees have pledged shares to charitable organizations of their choice. The company will add its own shares to these donations: at a 3-to-1 ratio for longer-serving employees and 1-to-1 for new employees. In its S-1 filings, Anthropic said it had reserved more than 550 million shares to fulfill commitments made before September 30.
Based on the number of reserved shares and the company’s disclosure that 84% of them are intended to match donations from early employees, TechCrunch estimates the employees’ pledged donations at $23 billion. Anthropic’s contribution would then amount to about $53 billion; the final figure will depend on the share price when the shares are transferred.
Together, the donations and company contributions could total $156 billion. This estimate assumes employees’ pledged donations are actually transferred to charitable organizations.
Where the money could go
Interviews with employees and organizations suggest that donors are likely to support charitable projects associated with effective altruism. TechCrunch names Longview Philanthropy, GiveWell, Coefficient Giving, Foresight Institute and Giving What We Can as possible recipients. A significant share of the money is likely to go to AI safety research, a concern for many Anthropic employees.
The ultimate impact depends on whether employees fulfill their pledges and when Anthropic transfers the shares, since the timing will affect the value of its contribution. The shares may also go to donor-advised funds; in that case, the money is unlikely to reach its final recipients immediately. Dataist’s take: $156 billion is a potential scale, not a guaranteed amount of charitable aid.
Daily AI news
Every day we pick what actually matters in AI and explain it plainly — no hype, no filler. Subscribe if you want to follow where the industry is going.
Only what matters — every day
Follow on X