Anthropic's draft IPO filing shows $80 billion loss and $750 billion compute plan
Archived — this story has rotated out of today’s deck. It is kept here in full.
The gist
Anthropic's draft S-1 showed a first-half 2026 net loss above $80 billion on $16.3 billion revenue.
The filing gives the first look at its finances before a mega-IPO.
Background
Anthropic is a San Francisco-based artificial intelligence company, maker of the Claude AI models, preparing a large IPO. It confidentially submitted a draft S-1 (the registration statement companies file before going public) to the SEC in early June 2026, joining a race of mega-AI IPO candidates. The Wall Street Journal's Oct. 20 report, citing the draft prospectus, gave the first detailed look inside the filing, showing steep losses alongside fast revenue growth and huge planned spending on computing capacity.
How it unfolded
- Jun 1, 2026Anthropic confidentially submitted its draft S-1 to the SEC, as it announced at the time.
- Oct 14, 2026New York City Comptroller Mark Levine sent a letter to CEO Dario Amodei urging the company to delay its IPO and focus on AI safety.
- Oct 19, 2026On prediction market Kalshi, Anthropic had 75% odds to officially announce an IPO this year, the only one of 23 listed companies above 50%.
- Oct 20, 2026The Wall Street Journal reported, citing the draft prospectus, a first-half 2026 net loss of more than $80 billion, including $72 billion in accounting charges, versus $13 billion a year earlier; revenue was $16.3 billion, up from $1.1 billion.
- Oct 20, 2026The draft indicated Anthropic could spend $750 billion on additional computing capacity in coming years, including roughly $450 billion in cloud compute agreements over the next decade.
- Oct 20, 2026Bloomberg reported Anthropic is expected to file publicly in the coming weeks and begin marketing the IPO as soon as the week of November 9.
- Oct 21, 2026The public version of the S-1 was not yet on the SEC's EDGAR database.
Who’s saying what
- Official
- NYC Comptroller Mark Levine urged Anthropic to delay its IPO, saying it should first address AI safety concerns and establish a governance regime, and that preparing for an offering diverts resources from that work.
- Analysts
- 22V Research analyst Dauvin Peterson said more questions than answers emerged around slowing revenue growth in September and whether Anthropic can grow as the market expects amid large language model pricing pressure.
- Caution
- Former Anthropic researcher Jacob Coxon claimed insiders believe AI could kill us all by the end of the decade.
Still unverified
The financial figures come from a draft prospectus obtained by the Wall Street Journal; the public version was not yet on SEC EDGAR as of Oct. 21. The $750 billion compute figure includes deals Anthropic is still negotiating. Jacob Coxon's claim about insiders' AI views is cited as his claim.