Analysis: Anthropic's Revenue Growth Slowed in the Third Quarter
Anthropic's third-quarter revenue reportedly reached $16.8 billion, but the increase was smaller than in the second quarter. An analyst concludes that growth nearly stalled between July and September.
Anthropic, the AI company that develops the Claude language models, reported third-quarter revenue of $16.8 billion, according to The New York Times, which cited figures the company shared with investors. An estimate published on the analytical blog Regenerator suggests that the pace of growth has slowed markedly. That matters because the trend bears on the valuation of the company's expected stock market debut and on investor expectations for the AI sector as a whole.
Quarterly figures
Based on the published data, the company's revenue this year was:
- first quarter: $4.8 billion;
- second quarter: $11.6 billion, up $6.8 billion, or 141%, from the previous quarter;
- third quarter: $16.8 billion, up $5.2 billion, or 45%.
In absolute terms, revenue therefore rose by more in the second quarter than in the third.
Monthly estimates
The author puts average monthly revenue growth at roughly $760 million in the second quarter and $580 million in the third. Taking into account the reported $65 billion annualized revenue run-rate in July, which equates to about $5.4 billion of monthly revenue, he infers that growth in August and September was only modest. According to the DealBook newsletter, the company did not give an annualized figure based on September revenue at the investor meeting; the analyst reads this as a sign of weaker growth.
Possible explanations
The author lists several possibilities: seasonality, since August is typically a slow month; the company being unable to keep up with demand; or a moderation in AI spending growth, including because businesses are using cheaper Chinese models for some tasks. In his view, usage data points to the last of these.
He argues this could also affect so-called neocloud providers such as CoreWeave, which have borrowed tens of billions of dollars on the assumption of rapid growth in demand for computing capacity. The monthly figures are the author's rough estimates, not company-confirmed data.
