Anthropic's 1,100% revenue jump meets an $895B bill, making its IPO a referendum on AI's economics

Anthropic reported third-quarter revenue of $17 billion, up from $1.4 billion a year earlier, according to AFP. That is an increase of more than 1,100%. The figures come from an amended filing ahead of its Wall Street listing later this month. Its net loss for July to September was $23.3 billion. The filing says this came mostly from non-cash accounting charges tied to the company's rising valuation. Its operating loss was $567 million, which the company linked to paying more for short-term computing capacity. Anthropic has also committed to spend about $895 billion, mostly on compute, with $70 billion due before September 2027.

The scale of the listing is unprecedented. Several outlets report that Anthropic aims to raise around $100 billion at a valuation of $1.5 trillion to $2 trillion. That would beat the record set in June by SpaceX, which raised $86 billion. The roughly 400-page filing spends more than 90 pages on risks. They include the recent reports of AI systems behaving in unexpected and harmful ways. The structure is unusual too. Anthropic is a public benefit corporation, a legal form that requires it to weigh its safety mission alongside shareholder returns. Its co-founders keep 50.1% of the votes through a single special share, which leaves public investors little say over the company's direction.

The backdrop is tense. AFP notes that revenue growth slowed compared with earlier quarters, amid fierce competition from OpenAI and from cheaper models, many of them Chinese. Last week, worries about AI spending triggered a sharp sell-off in big tech stocks. Meta, Alphabet and Oracle, which is pushing ahead with a $20 billion plan, are also turning to markets to fund their buildouts.

Our reading: this IPO is a referendum on AI's unit economics, not just on one company. Two facts can be true at once. Anthropic's demand is real, and growing more than twelvefold in a year is not hype. Its cost base is also enormous, and competition from cheaper open models, many of them Chinese, is narrowing its margins. That fits our view that China is closing the gap fastest in open-weight models. The governance terms deserve as much attention as the revenue. Founder control and a mission-bound corporate form could shield safety commitments from short-term pressure. They could also shield management from accountability. Investors are effectively being asked to trust the mission. Our position remains nuanced optimism. This capital will fund the compute behind AI's long-term promise in science and health. The $70 billion due within a year, however, is the near-term test to watch.