AI Labs & Models

Anthropic IPO filing reveals high margins and Q3 operating loss

Anthropic's S1 filing shows an 85% gross margin and significant R&D spending, but a Q3 operating loss driven by compute costs and donations. Investors are weighing these figures against the potential impact of recursive self-improvement on AI valuations.

· 3 sources · Anthropic financials

Anthropic's S1 filing confirms audited unit economics, showing September revenue growing from $1.4 billion last year to $16.8 billion this year. The company reported an 85% gross margin on cost of sales, comparable to Google's 87-88% at its IPO and Meta's 77-78% in 2011. R&D costs, primarily for building the next model, account for 59.5% of revenue in the first half, compared to 9.5% for Google and 10.5% for Meta. Combined, 75% of revenue goes to running existing models and building new ones, with training spend rising from $4.5 billion in the second half of last year to $9.7 billion in the first half of this year.

Despite these margins, Anthropic updated its IPO documents to reveal an operating loss of over $500 million in the third quarter, reversing previous profitability. The loss was driven by a shortage of compute resources, which forced the company to purchase expensive short-term computing power, and a donation-matching program that resulted in approximately $2 billion in costs during the quarter. The market reaction was muted, with no significant new concerns raised about the AI sector's financial health. Meanwhile, investors are analyzing the potential impact of recursive self-improvement (RSI) on valuations. Anthropic co-founder Jack Clark stated that RSI might be the most important technical trend in the world, while Alphabet CEO Sundar Pichai predicted continued acceleration in the next 12 months due to RSI. Anthropic's responsible scaling policy notes it does not anticipate safely using Claude to fully autonomously design successive frontier models without meaningful human supervision.