Anthropic's founders want 50.1% of the votes on a $2T listing — but the board answers to outside trustees holding zero equity

The headline analogy is misleading. Palantir gives its three co-founders — Alex Karp, Stephen Cohen and Peter Thiel — Class F shares deposited into a Founder Voting Trust that delivers49.999999% of the votes, according to the company's S-1 filing with the SEC. That is a defense: no two of the three can be outvoted by every outside shareholder combined. Anthropic's ask is different. 50.1% is the minimum needed to win a shareholder vote against a united opposition. The new class of stock carries voting power without proportionate economic rights, and the threshold holds as long as three of the seven co-founders continue to hold their stipulated number of shares — meaning the founders can lose a majority of themselves and still control the company.

The founders' vote does not pick the board. Anthropic's Long-Term Benefit Trust — five members holding zero equity, no dividends, no capital appreciation — currently has the right to elect a majority of directors through Class T shares, per Anthropic governance research compiled by Klover AI. One named trustee is former California Supreme Court Justice Mariano-Florentino \"Tino\" Cuéllar. Under the IPO structure, the founders' slate of director nominees expands from two to three, while the trust retains its majority — meaning four of seven board seats answer to outsiders holding no equity.

Anthropic told prospective investors it generated more than $11.5B of revenue in Q2 2026 — up from $787M a year earlier — with positive adjusted operating income, according to Bloomberg's August 14, 2026 report and CNBC's same-day coverage. Q1 came in at $4.73B, putting annualized run-rate revenue at roughly $46B by mid-year and the company on track for $65B-plus on management's May 28, 2026 Series H announcement ($65B raised at a $965B post-money valuation, led by Altimeter, Dragoneer, Greenoaks and Sequoia). At a $2T target valuation, public-market investors are paying about 43x annualized revenue — a premium to every large-cap tech comparable.

Anthropic has already locked in the upstream capacity. The September 24 voting-control news is the last piece of an18-month build-out: $13B of cumulative equity from Amazon ($5B in April 2026 plus up to $20B tied to milestones, per the Anthropic–Amazon announcement), up to $40B from Alphabet in cash and compute per TechCrunch, and a 3.5-gigawatt TPU commitment routed through Broadcom per the Anthropic–Google–Broadcom partnership release. A successful IPO validates that compute supply stack — and the listed names inside it.

Founder-led dual-class structures are now standard at US mega-cap listings. Meta listed with two classes in 2012; Snap went further in 2017 by issuing only non-voting shares to the public; SpaceX's 2026 IPO filing carries Musk-controlled Class B shares with disproportionate votes per Reuters' dual-class explainer. Anthropic's twist is that the Long-Term Benefit Trust — not the public shareholders and not the investors who poured in $25B from Amazon and $40B from Alphabet — appoints the majority of the board. The trust can also remove directors whose behavior conflicts with responsible AI development.

The right comparison is not Palantir. Palantir's Class F is a defensive structure — it stops the founders from being outvoted. Anthropic's 50.1% is offensive: the founders and the trust together can pass any resolution over the objection of every other shareholder, including Amazon and Google. For investors underwriting a $2T float, that is the price — and it is the precedent that every other frontier-AI IPO will inherit.