Artificial intelligence company Anthropic saw its revenue surge by 45% to $16.8bn (£12.7bn) in the three months to the end of September, the BBC understands.
The figures were shared at a briefing for prospective investors at the company's San Francisco headquarters on Wednesday, as the maker of the Claude chatbot lays the groundwork for a stock market listing expected before the end of the year.
Anthropic also reportedly told investors it had made a profit for the second quarter in a row. A company spokesperson declined to comment.
A steep climb
Anthropic's revenue has risen every quarter this year, from $4.7bn in the first three months to $11.6bn in the second and now $16.8bn.
Its figures, along with those of rival OpenAI, are closely watched as a gauge of whether the AI industry can generate enough money to pay for the vast data centres and chips it needs.
Markets take fright
The numbers emerged at a nervous moment for technology stocks.
On Wednesday, shares in US chipmakers including AMD, Nvidia and Intel fell sharply. On Thursday morning the sell-off spread to Asia. Japanese tech investor SoftBank fell by more than 9%. South Korean semiconductor giants Samsung and SK Hynix dropped 9.4% and 10.8% respectively.
Trading on South Korea's Kospi and Kosdaq exchanges was temporarily halted after falls of more than 8%, while Japan's Nikkei 225 fell by more than 5%.
The slide was triggered in part by comments from Hock Tan, chief executive of US chip firm Broadcom, at a conference in Singapore on Monday. He said many investors had "no clue" and had set "extreme expectations only to find out that it's almost like they are in a casino".
The falls deepened after US inflation data on Wednesday showed prices rising faster than expected in September, raising fears that the Federal Reserve could increase interest rates – bad news for companies whose valuations depend on profits far in the future.
Growth versus bubble
Anthropic's results cut both ways for the market. On one hand, they show that businesses and consumers are paying real money for AI services at a remarkable rate. On the other, investors are asking whether the hundreds of billions of dollars being spent on infrastructure can ever be earned back across the industry.
A successful Anthropic listing would be one of the biggest technology flotations in years and a crucial test of public investors' appetite for AI.
The safety question
There are also concerns about the safety of AI systems. OpenAI chief executive Sam Altman said in September that his company would not go public this year while it addressed concerns about the technology's safety. The ChatGPT developer is expected eventually to seek a listing valuing it at more than $1tn.
Anthropic has positioned itself as a safety-focused developer. Its chief executive, Dario Amodei, has called for the pace of AI development to slow down – an unusual message from a company racing to raise money from public markets.
Analysts note that sell-offs of this kind have often hit hardware makers hardest, because their revenues depend on AI developers continuing to spend at record levels. Any hint that the biggest buyers of chips might slow their orders tends to ripple instantly through Seoul, Tokyo and Taipei, where the semiconductor industry dominates the stock market.
Expanding abroad
The company continues to grow internationally and opened an office in Singapore on Wednesday.
For the UK, which has courted AI firms with promises of investment zones and computing capacity, the health of companies such as Anthropic matters directly: their spending decisions shape where data centres are built and where skilled jobs are created.
For now, Anthropic has a remarkable story to tell investors. The question is whether, after this week's sell-off, they are in the mood to listen.
