The $965 Billion AI Company You Cannot Legally Sue

When the S-1 filing landed on June 1, 2026, it showed what any IPO filing shows — revenue, valuation, share structure. But beneath those numbers lies something no public market has ever encountered at this scale: a company whose shareholders are legally prohibited from forcing it to maximize their returns.

Anthropic filed a confidential draft S-1 with the SEC on June 1, revealing $47 billion in annualized revenue run-rate and a $65 billion Series H closed just weeks earlier at a $965 billion post-money valuation — the largest private funding round in history. The company, founded in 2021 by Dario Amodei, Daniela Amodei, and other former OpenAI researchers, has compute agreements covering more than 10 gigawatts with Amazon, Google, Broadcom, and SpaceX. Claude runs on AWS, Google Cloud, and Microsoft Azure. CFO Krishna Rao called the moment historic demand.

He wasn't exaggerating. But the most consequential fact in that S-1 isn't the revenue number.

What the Trust Actually Means

Anthropic is structured as a public benefit corporation. That's not unusual. What makes Anthropic different is the Long-Term Benefit Trust: a Delaware purpose trust with no beneficiaries. Under the structure documented in Anthropic's governance filings and analyzed by Yale Law School professor John Morley and Wilson Sonsini partners David Berger and Amy Simmerman, the Trust is managed for the achievement of a purpose — Anthropic's mission — rather than for the financial benefit of any identified party.

In a conventional trust, beneficiaries can sue trustees who fail to act in their interests. The Long-Term Benefit Trust has no such beneficiaries. No public shareholder can walk into a courtroom and argue that the board is insufficiently focused on share price. The fiduciary duty, such as it is, runs to the mission, not to the investor.

The Trust's voting power scales as Anthropic raises more capital: it starts at one-fifth of board seats and grows toward a majority as equity holders accumulate. Class T shares give trustees this escalating control. A sophisticated set of pre-IPO investors — including Altimeter Capital, Dragoneer, Greenoaks, Sequoia, GIC, Coatue, Baillie Gifford, Blackstone, Fidelity, T. Rowe Price, and Temasek — accepted those terms and signed on anyway.

That was their choice. Public retail investors have no such negotiation.

The Three-Way Race

Anthropic filed behind SpaceX, which submitted its public S-1 on May 20 and is targeting a June 12 Nasdaq listing under ticker SPCX at a $1.75–1.8 trillion valuation, seeking up to $75 billion in the largest IPO in history. OpenAI confidentially filed around May 22 and targets September 2026 at a $1 trillion-plus valuation.

Anthropic sits between them: more established as a standalone AI business than xAI, which burned $14 billion in cash against $3.2 billion in revenue in 2025, but racing in the same direction. Unlike SpaceX, which has Starlink generating $11.4 billion in revenue and $4.4 billion in operating income, Anthropic has no obvious self-funding profit center. What it has is $47 billion in annualized revenue, growing.

The question isn't whether Anthropic can go public. The revenue run-rate answers that. The question is what public investors are actually buying.

The Transparency Paradox

Here is the irony the S-1 crystallizes: Anthropic's IPO may produce the most rigorously transparent AI company in history, by one definition. Public companies disclose material information to the SEC continuously. Every quarter, Anthropic will file financial statements, explain material risks, and describe material transactions in terms the SEC can examine and shareholders can challenge.

But the governance structure means that one category of information is permanently off-limits to shareholder challenge: the company's decisions about when to slow down, what safety research to prioritize, and whether to pass on a commercially valuable deployment because the risk calculus doesn't pencil out. Those decisions answer to the Trust, not to investors. And the Trust cannot be sued for making them.

This is not opacity in the traditional sense. Anthropic will disclose plenty. It will likely describe its safety research, its risk frameworks, its governance processes — because PBics have disclosure obligations around their specific public benefit commitments. Readers who want to understand how Anthropic thinks about catastrophic risk will have more documented information than they have about any comparable private lab.

What shareholders cannot do is compel the company to act on that information in any particular direction. Transparency about process coexists with permanent insulation from accountability for outcomes.

The Price of Safety

The $965 billion valuation implies roughly a 20x multiple on annualized revenue — a premium that requires sustained triple-digit growth to justify on conventional metrics. Investors pricing that multiple are implicitly betting that the governance structure will not meaningfully impede commercial execution. They may be right. The Amodeis have run the company under these constraints for five years and built $47 billion in annualized revenue. Sophisticated institutional investors have held those same terms since the Series C.

But the structure has never been tested at public market scale, under the pressure of quarterly earnings calls, activist hedge funds, and institutional investors with 13F filing requirements and proxy voting obligations. The Trust's escalating control is designed to keep the mission intact as outside capital grows. Whether it does so when a trillion-dollar public market starts asking uncomfortable questions about opportunity cost is a different proposition.

SpaceX has Elon Musk's control as an anchor — his 85% voting power makes outside pressure structurally irrelevant. Anthropic has the Trust. Both structures immunize the company from shareholder primacy. Anthropic's version is more legible, more deliberately designed, and less dependent on a single individual's incentives. Whether that makes it more durable or just more novel is something the public markets are about to find out.

Anthropic, advised by Wilson Sonsini Goodrich & Rosati, has not disclosed the timing of its listing. The confidential S-1 will become public in the weeks ahead, and with it, the full governance disclosure. Investors who buy in will be buying a stake in a company that answers to its mission — and to nobody who can sue when the mission and the share price diverge.