2024-09-25
한어Русский языкEnglishFrançaisIndonesianSanskrit日本語DeutschPortuguêsΕλληνικάespañolItalianoSuomalainenLatina
author: li dan
source: hard ai
some newly disclosed revenue data is proving that business is booming for hot artificial intelligence (ai) startups and that the market's high valuations for these unicorns are unjustified.
on tuesday, september 24, eastern time, cnbc reported that documents obtained by its reporters showed that anthropic, a us startup that developed the large language model (llm) claude and is a rival of openai, disclosed to investors that it expects to reach $1 billion in revenue this year. that would mean that anthropic's revenue would increase by 1,100% over the previous year.
the report mentioned that anthropic also disclosed that in this year's revenue, sales from third-party application programming interfaces (apis) accounted for 60% to 75%, that is, the relevant revenue for developers and third parties such as amazon cloud aws to call anthropic models accounted for as much as three-quarters, direct api sales accounted for 10% to 25%, chat tool subscriptions accounted for 15%, and professional services accounted for 2%.
also on tuesday, technology media the information reported that helping giants such as google and meta improve their llm accuracy is becoming a big business. according to insiders, scale ai, an ai company that provides data annotation services for ai model training, had revenue of nearly $400 million in the first half of this year, nearly three times the same period last year.
the report said that although scale ai has not turned a profit in eight years, it has increased its operating profit margin this year. in the first half of last year, scale ai spent $1.50 for every $1 in revenue it generated, and in the first half of this year, the average cost per $1 in revenue dropped to about $1.20. if only the direct costs of the business are considered, scale ai only has about half of its revenue invested in expenses, that is, the gross profit margin is slightly less than 50%, which is a decrease from about 57% in the first half of 2022.
the above reported data shows that scale ai's gross margin and revenue are slightly lower than the forecast level released by the company to potential investors earlier this year. at that time, the company predicted sales of about $415 million in the first half of the year and a gross margin of about 51%, which was a few percentage points higher than the actual reported figures.
coincidentally, the two ai startups mentioned above that have reported a surge in revenue both have high valuations of tens of billions of dollars.
this mondaymentioned that media reports cited existing investors as saying that anthropic has begun discussing raising funds with investors. the deal could value the company at $30 billion to $40 billion, compared with a round of financing completed earlier this year.financingthe report did not reveal which investors would participate, suggesting that it could be corporate investors as before.
anthropic may eventually hire an investment bank to help facilitate a new round of financing, according to the report. startups typically don’t hire investment banks to solicit investments unless the financing process is complex, such as involvingsovereign wealth fundsor corporate investors. the aforementioned investors said that anthropic has been raising funds from corporate investors and is likely to do so again. in fact, amazon announced in march this year that it would invest an additional $2.75 billion in anthropic on top of the $1.25 billion investment announced in september last year.
scale ai just completed a round of financing in the first half of this year. in may, scale ai completed a $1 billion series f financing, led by accel, which led the company's series a financing. the valuation thus reached $13.8 billion. in addition to existing investors such as amazon, meta, nvidia, tiger global management, coatue, y combinator, index ventures, and founders fund, this round of financing also attracted new investors, such as cisco, intel, and amd's respective venture capital institutions.