Three Stanford Physicists and an S-1: The Arithmetic Behind Anthropic’s Filing

Start with the deal, because the deal is enormous. On Monday, Anthropic filed its registration statement on Form S-1 with the Securities and Exchange Commission, proposing to list on Nasdaq under the ticker “ANT,” with Goldman Sachs and Morgan Stanley leading the underwriting. The filing doesn’t yet give a share count or a price range. It does, however, contain a fact that Stanford readers will find more interesting than any multiple: three of the company’s seven co-founders trained in physics on the Farm.

The Stanford line in the prospectus

The management biographies in the S-1 are unusually plain about it. Dario Amodei, the chief executive, “holds a Ph.D. in Biophysics from Princeton University and a B.S. in Physics from Stanford University,” and also completed a postdoctoral fellowship at the Stanford University School of Medicine before going on to Google Brain and then OpenAI, where he rose to vice president of research. Jared Kaplan, the chief science officer, holds a Stanford bachelor’s degree in physics and mathematics and a Harvard Ph.D. in physics. The filing calls him “one of the principal authors of the neural scaling laws that have shaped modern LLM development.” Sam McCandlish, the chief architect, went the other way round, with a Brandeis bachelor’s degree and a Ph.D. in physics from Stanford. At OpenAI he led a team whose scaling-law work, the prospectus says, “helped pave the way” for GPT-3.

The boardroom has a Stanford connection too. Reed Hastings, the Netflix co-founder, has been a director since May 2025, and the S-1 lists his M.S.C.S. in artificial intelligence from Stanford. That makes four Stanford degrees at the top of a company that the S-1 shows growing faster than almost anything the public markets have been asked to price.

It is worth pausing on what the three founders have in common. Scaling laws, the idea that a model’s performance improves predictably as a power-law function of its size, its data and its compute, are a physicist’s way of looking at a machine-learning problem. Both of the founders the filing credits with that line of work did their physics at Stanford. The scaling-law bet, in other words, is the company’s founding thesis, and the S-1 is the first audited test of what it is worth.

The arithmetic

Here are the reported numbers. Anthropic’s revenue was $41 million in 2023, $386 million in 2024 and $4.6 billion in 2025. For the six months ended June 30, 2026, it was $16.3 billion, against $1.2 billion in the same period a year earlier, which the company calls growth of 1,300 percent. Across two fiscal years, from 2023 to 2025, that is roughly a 112-fold increase.

Losses are the other half of the ledger. The operating loss narrowed to $693 million in the first half of 2026, from $2.1 billion a year earlier. The headline net loss is far larger, about $80.7 billion for the half. According to the filing, most of that, about $72.4 billion, comes from changes in the fair value of convertible notes, derivatives and SAFEs. That is an accounting charge driven by the company’s own soaring valuation rather than cash leaving the building. Readers should keep the two figures apart.

Now the valuation, where reported figures have to be kept separate from estimates. CNBC’s coverage notes that Anthropic’s last private round, in May, valued it at $965 billion. That is a reported figure. As a back-of-envelope comparison, and nothing more, doubling the first-half revenue gives an annualized run rate of about $32.6 billion, which puts the May valuation near 30 times run-rate sales. Estimates for the public debut run higher. MIT’s Sinan Aral told CNBC he could see a “$1.5 to $2 trillion valuation.” That is one academic’s estimate, not a number in the filing, and the filing itself warns: “We do not expect our revenue will continue to grow at comparable rates, and we cannot assure you that it will continue to grow at all.”

For the venture class, the story is in the step-ups. Menlo Ventures’ Matt Murphy told CNBC the firm first invested at a $4 billion valuation. Measured against the $965 billion May mark, that is roughly a 240-fold paper gain on the early shares, before dilution, and it is the kind of number that will be quoted at Sand Hill Road dinners for a decade. TechCrunch reports that more than 11,900 customers now spend over $100,000 a year with Anthropic, about 1,900 spend over $1 million, and roughly 250 spend over $10 million.

Control, and conscience, by design

A physicist’s company has, fittingly, filed an unusually engineered capital structure. Anthropic is a Delaware public benefit corporation, and its stated purpose is “the responsible development and maintenance of advanced AI for the long-term benefit of humanity.” A single share of Class F stock, which carries no economic rights, will be held by a Founder LLC whose only members at first are the seven co-founders. That share carries 50.1 percent of the voting power on most matters. Separately, a Long-Term Benefit Trust of financially disinterested trustees can elect and remove a majority of the board. TechCrunch notes that the arrangement has already drawn criticism because voting power does not track economic ownership.

There is a philanthropic layer as well. The company has contributed 846,840,960 Class N shares, 5 percent of its fully diluted shares as of June 30, to a wholly owned Anthropic Public Benefit Fund. Each co-founder, including the three Stanford physicists, has pledged to dedicate 80 percent of their personal equity to philanthropic causes. The S-1 is candid that these pledges are “voluntary, revocable expressions” of intent and not binding obligations. Even so, if the pledges are honored at anything like today’s valuations, Stanford alumni will be directing philanthropy at a scale the university’s own donor rolls have rarely seen.

The next cohort is already raising

Every great listing raises the question of who comes next, and the answer, again, runs through Stanford. Last week TechCrunch reported that Flapping Airplanes, a lab about a year old, raised a $1 billion Series A co-led by Index and Lux at a reported $6 billion valuation. Nine months earlier it had raised a $180 million seed round at $1.5 billion, so the valuation has quadrupled. Two of its three founders are brothers with Stanford doctoral ties: Asher Spector earned his Ph.D. in statistics at Stanford with a dissertation advised by Emmanuel Candès, and Ben Spector was a computer science Ph.D. student there.

What makes the pairing instructive is the thesis. Anthropic’s founders made their names showing that more data and more compute buy more capability. Flapping Airplanes is betting on the reverse: training capable models “without ingesting half the internet.” “A model that’s a million times more data efficient is probably a million times easier to put into the economy,” Asher Spector told TechCrunch. One generation of Stanford-trained scientists is about to test the public markets with the scaling thesis. The next is raising billions to argue that the thesis has limits.

The pricing, the roadshow and the lock-ups lie ahead, and so do the regulatory and political fights that have shadowed the company this year. Nothing here is a recommendation to buy or sell anything. What is already on the public record is a 1,300 percent growth rate, a $965 billion private mark and a governance structure drawn up by people who think in power laws. Part of that story began in the physics classrooms of the Farm, and Stanford would do well to watch how it ends.

Photo: TechCrunch / Wikimedia Commons, CC BY 2.0