Anthropic Just Pledged 5% to a Public Benefit Fund — and 50.1% of the Vote to Its Founders

Anthropic's Sunday announcement frames the Public Benefit Fund as a structural commitment, not a one-time donation: subject to stockholder approval, 5% of Anthropic's shares will be committed today, with an additional 2% added at the board's discretion for every 50% increase in the company's value above a set baseline. The company will not say what the baseline is yet — that's reserved for a forthcoming public benefit report — but the mechanism is the article's load-bearing piece.

The structure matters more than the percentage. A 501(c)(3) would lock the vehicle into U.S. nonprofit rules and force grantmaking; a wholly-owned LLC gives the board optionality to pursue \"large donations to traditional nonprofits or direct distributions to specific populations\" — language Anthropic used to telegraph what the money could ultimately fund. The company cited labor-market displacement as the first concrete example: it has already committed $200 million to its Economic Futures Program and an Economic Futures Research Fund, and is now positioning the new LLC to go bigger.

Anthropic is explicit that the headline 5% is \"instead of escalating compensation awards to their executives\" — equity that would otherwise accrue to insiders is being rerouted into the fund. The math implies a deliberate trade: founders give up some equity upside, and in return the company locks in a public-benefit claim that compounds with valuation rather than with executive retention.

The Public Benefit Fund isn't the only governance item on the proxy. Roughly three weeks earlier, Anthropic filed paperwork asking shareholders to approve a Founder LLC that would give CEO Dario Amodei and his six co-founders 50.1% of voting power on most corporate matters through a single Class F share. The structure applies \"as long as at least three of the seven co-founders retain a minimum number of shares,\" with one notable carve-out: the founders' voting bloc does not extend to board member elections.

The framing inside the company is that founder voting power is the price of mission alignment — without it, the PBC's public-benefit language could be overridden by future shareholder pressure. To an outside investor, however, it looks like dual-class on steroids: a non-economic voting share held by a private LLC, sitting alongside a 5% public-benefit LLC that also holds economic equity. Both vote items travel together; rejecting one likely kills the other.

Before the Public Benefit Fund or the Founder LLC existed, Anthropic built a third body: the Long-Term Benefit Trust (LTBT), an independent five-trustee committee that holds Class T shares worth $0 — no equity, no dividends, no liquidation preference — but elects and removes Anthropic's board members. As of April 2026, the LTBT-appointed directors held three of six seated board seats, a working majority.

Anthropic's Ben Bernanke appointment reads as mission-flavored talent acquisition: the company explicitly cited his \"economic impacts\" expertise when adding him to the LTBT in July 2026. Combined with the new Public Benefit Fund, Anthropic now has three independent mission-aligned bodies — the LTBT (board appointments), the new LLC (economic grants), and the Founder LLC (founder voting) — each of which serves as a check on the others. Whether that creates stable governance or factional capture is the open question for IPO investors.

Anthropic's reported IPO target of more than $2T — more than double the $965B Series H post-money set in May 2026 — implies roughly 10x projected 2028 revenue (~$190–200B internal forecast) and about 31x the late-July 2026 run rate of ~$65B. The implied multiple isn't defensible on near-term earnings; it depends on a multi-year compute-cost trajectory bending in Anthropic's favor. The Public Benefit Fund's valuation-linked escalator converts that bet into a public-good claim — if the $2T math holds, the LLC's stake could climb meaningfully above 5%.

Anthropic is the rare AI lab whose customer base and capex base are the same three names: Amazon, Alphabet, and Microsoft. Roughly 47% of Anthropic's 2025 revenue flowed through Amazon and Google cloud marketplaces, and Barclays estimates that Anthropic plus OpenAI together account for at least 13% of Amazon's AWS revenue in 2026 — about $23 billion. That makes the IPO a derivative trade on the largest hyperscalers' revenue mix, not just an AI sentiment trade.

Amazon already booked a $53.4B accounting gain on its Anthropic stake in Q2 2026 alone — a real reported number, not a revaluation. If Anthropic prices above the Series H mark, Alphabet and Microsoft sit on similar paper gains, with Microsoft's $5B deal only seven months old. The Public Benefit Fund announcement doesn't change those accounting marks, but it makes the post-IPO ownership story cleaner: ~20% earmarked for public benefit, 50.1% of the vote with founders, the residual — about 75–80% of the economic equity — with existing investors and the public float.

The interesting supply-chain shift is the mix. With Anthropic's Q3 disclosure showing the Google TPU angle working — Claude training increasingly on TPU capacity — Broadcom becomes the most leveraged public beneficiary. NVIDIA doesn't lose revenue (Azure commitments keep growing) but cedes share inside Anthropic's stack. SK Hynix and TSMC are utility plays: both win regardless of which accelerator architecture wins, because both architectures depend on the same HBM and foundry capacity.

Three dates matter more than the Oct. 18 announcement itself. First, the stockholder vote on both the Public Benefit Fund and the Founder LLC — shareholders must approve both the 5% commitment and the 50.1% voting structure; rejection of either likely forces renegotiation. Second, the public S-1 filing — currently 130+ days past the June 1, 2026 confidential submission; the public filing is required at least 15 days before institutional marketing begins. Third, the Nasdaq listing, currently targeted before Thanksgiving (Nov. 26, 2026).

The real question isn't whether Anthropic lists — it will — but whether the structure scales. The Public Benefit Fund's board-discretion escalator means the LLC's stake grows automatically with valuation, but the LTBT trustees must still approve how that equity gets spent. That creates an unusual bottleneck: the same five trustees (now including a former Fed chair) sit across both the board and the fund's eventual grantmaking decisions, while the Founder LLC's 50.1% voting share keeps the founders in the room on any structural change. If it works, Anthropic becomes the template for every future AI lab listing; if the bodies deadlock, expect a 2027 governance crisis that rerates the entire 'mission-aligned IPO' category.