Analysis
Anthropic said Oct. 18 it will commit 5% of its shares to a newly created Anthropic Public Benefit Fund, a wholly-owned entity the company says will give it "increased flexibility and scale" to fund projects addressing AI's effect on society, according to TechCrunch. The pledge is subject to stockholder approval, and Anthropic's board can add another 2% of shares for every 50% increase in the company's value above an unspecified baseline -- meaning the stake grows automatically alongside the valuation gains Anthropic is chasing into its expected IPO.
The fund extends a pattern Anthropic set at its founding. The AI safety company, which organized as a Delaware public benefit corporation in 2021 under founders Dario Amodei, Daniela Amodei, Chris Olah and Jared Kaplan, already runs a Long-Term Benefit Trust that elects a growing share of its board seats, and in 2026 it committed $200 million to study AI's impact on the labor market.
Bloomberg reported the new fund arrives as Anthropic prepares for what could be one of the largest IPOs ever. The company's valuation has climbed fast this year, from its February Series G through May's Series H -- and reported IPO targets since then have ranged as high as $2 trillion.
## What the headline misses Other AI labs have made softer public-benefit commitments: OpenAI restructured into a capped-profit model under a nonprofit parent, while xAI and Google DeepMind have no comparable equity carve-out. On the older-economy end, Patagonia's 2022 move to give 98% of its stock to a trust and a nonprofit is the closest precedent for a founder-controlled company diverting equity at scale -- though Anthropic's structure, a for-profit LLC it fully owns rather than an independent trust or 501(c)(3), keeps the money and the decision-making inside the company rather than outside it. Anthropic also said it will share more about the fund's structure and governance "in the coming months," meaning the pledge, for now, is a number without a disclosed mechanism.
Five percent of a company valued near $965 billion is roughly $48 billion on paper, though the fund's actual payout depends on how and when Anthropic liquidates or distributes the shares.
Layered on top of the Long-Term Benefit Trust's governance stake and the earlier labor-market commitment, some outlets estimate Anthropic's total public-purpose equity commitments now approach a fifth of the company -- a scale with no clean comparison among venture-backed AI companies heading toward IPOs this size.
For VCs and later-stage investors, the fund is a dilution question dressed up as a governance one: every share routed to the fund is a share that doesn't go to common stockholders when Anthropic eventually lists. Investors underwriting Anthropic's IPO will need to understand exactly how the fund's cap-table carve-out works before pricing the deal, and a skeptic's read is that announcing a public-benefit fund days ahead of an IPO that could value the company near $2 trillion is also a well-timed answer to regulators and critics asking what AI companies owe the public as they get richer.
What to watch: whether Anthropic discloses the fund's baseline valuation for the 2% kicker, who sits on its governing board, and whether its first grants go to AI-safety research or broader social programs. Anthropic's safety posture is already under scrutiny for a parallel move -- Pulse has tracked the company giving outside evaluators like METR deeper, employee-like access to its systems this same month, part of the same broader push to answer critics before regulators do.
