Anthropic Files for IPO: $16.3 Billion in First-Half Revenue, an $80.7 Billion Net Loss, and Talk of a $1.5–1.75 Trillion Valuation
One prospectus, two numbers on facing pages: "$16.3 billion in first-half revenue" and "an $80.7 billion net loss." Put together, that's the AI industry's most honest self-portrait right now.
Three key facts
Anthropic filed its S-1 on October 26, aiming to go public in November. The prospectus shows first-half 2026 revenue of $16.3 billion and a net loss of $80.7 billion; share count and price range are undisclosed. CoinDesk reports, citing the Financial Times, that investors expect a $1.5 trillion to $1.75 trillion valuation; Fox Business's Liz Claman says a November 19 or 20 pricing remains on track. Earlier reports suggested a raise of as much as $100 billion at a $2 trillion valuation.
The balance sheet is deep, but the spending outlook is climbing. As of June 30, the company held $112.4 billion in cash and equivalents plus $19.3 billion in short-term investments. But the filing explicitly says infrastructure investment is expected to increase significantly, putting pressure on future cash flows.
Enterprise growth is the hardest number in the document. By September 30, more than 1,900 customers had at least $1 million in trailing annual revenue, including about 250 above $10 million. Market reaction was immediate: chip stocks rose pre-market after the filing, with Nvidia up 1.4%, Broadcom 1.5%, and Micron 2.3%; on Hyperliquid, the ANTH contract spiked to about $1,980 around the filing, then fell roughly 6% to about $1,857 within minutes.
Wang Dou's Take
Start with that "$80.7 billion net loss." This isn't an operating loss — a company with $16.3 billion in half-year revenue can't lose five times its revenue just on salaries and compute. Losses of that magnitude typically come from stock-based compensation, fair-value changes on convertible instruments, and other paper losses. In other words, the number says more about "this company's valuation is rising fast" than "this company is being run badly." That's the first trap in reading any AI prospectus.
The second trap is the "$1.75 trillion" rumor. If it happens, it exceeds the GDP of nearly every country on earth. And what backs it: 1,900 million-dollar customers and 250 ten-million-dollar customers. Pretty numbers — but divide them into $1.75 trillion and you find the market is pricing each "ten-million-dollar customer" in the billions. This isn't buying today's revenue; it's buying the belief that AI wins.
The third point is the most telling: the same document says AI, "if managed well," could accelerate scientific innovation and bring cures for major diseases — and, in the risk factors, that advanced AI systems could pose catastrophic or existential risks to humanity. A company selling the thing while warning the thing might destroy us, with both paragraphs in one file, reviewed by one legal team. That isn't a contradiction. It's the industry's most honest self-description: we know it's dangerous, all of us know, and we're going to build it anyway — and take it public.