Anthropic Q3 Revenue Hits $17 Billion in Updated IPO Filing

Enterprise customers and efficiency

Enterprise clients drive the business. Anthropic counts more than 11,900 customers spending over $100,000, more than 1,900 above $1 million and about 250 above $10 million. In the third quarter, compute costs rose 382% against revenue growth of 1,127%, and revenue per employee exceeded $6.6 million, up from $1.4 million a year earlier.

The filing also says the list-price value of Claude's internal use grew more than 250-fold between January and September 2026. In April, that monthly value exceeded monthly labor costs excluding stock compensation. Anthropic presents this "closing the loop", with Claude running experiments, processing training data, automating evaluations and safety audits, and writing and reviewing code, as a competitive strength.

New risk factor on AI self-acceleration

The new addition is a risk factor on AI self-acceleration. Anthropic says future models may soon reach the "dramatically accelerated progress attributable to AI" threshold in its Responsible Scaling Policy. If capabilities outpace safety, alignment and security work, or if the pace cannot be predicted, it may be unable to assess or mitigate the resulting risks, raising the likelihood of safety incidents and regulatory intervention.

As it nears what it calls recursive self-improvement, the company says it has plans to monitor and reduce risks and is investing in safety techniques. It does not expect that fully autonomous design of next-generation frontier models by Claude, without substantive human oversight and understanding, could be done safely, even if technically possible.

Contracted spending, mostly compute and infrastructure, totals about $895.4 billion. About $447.5 billion falls due by the end of 2030, the rest from 2031 to 2045, and roughly $69.8 billion over the next 12 months. Anthropic plans to list on Nasdaq under the ticker ANT; share count, price range and listing date are not yet set.