Snorkel AI raises $350 million at a $3.5 billion value; its price per dollar of revenue fell about sixfold

Snorkel AI, a San Francisco company that builds training data and test environments for AI developers, said on Tuesday it raised $350 million at a $3.5 billion valuation. Insight Partners and S32 co-led the round, according to the company's press release, with existing investors including Addition, Greylock, Lightspeed, GV and Wells Fargo also taking part.

Reuters, which first reported the round, said the new price is nearly three times the $1.3 billion valuation from Snorkel's $100 million raise in May 2025. The company told Reuters its annualized revenue run-rate has passed $350 million, up from roughly $20 million a year earlier, and that it expects to be profitable this year.

The number behind the number: a cheaper price per dollar of sales

"Valuation nearly triples" sounds like a hotter bet. The revenue figures say the opposite about price. Using the numbers Snorkel gave Reuters:

• Then: a $1.3 billion valuation against a run-rate of about $20 million a year ago works out to roughly 65 times revenue.
• Now: $3.5 billion against more than $350 million is about 10 times revenue.

The two points are a few months apart (the 2025 round was in May, the $20 million figure is from about September), so the older multiple is approximate. The direction is not in doubt: revenue grew more than 17 times while the valuation grew less than 3 times. Investors in this round are paying a small fraction of what the last round paid for each dollar of sales. That can reflect caution about how durable the revenue is, since much of it comes from a small number of very large AI labs, as much as it reflects any enthusiasm.

One more detail worth noting: the revenue and profitability figures come from Reuters' reporting of what the company said. Snorkel's own press release does not state a revenue number.

What changed in the business

Snorkel was founded in 2019 by researchers from the Stanford AI Lab and first sold software for labeling data. In September 2025 it launched what it calls expert data-as-a-service, supplying finished datasets and reinforcement-learning environments directly to customers. Reuters said that business drove the jump in revenue. The company pairs human experts, who write tasks and grading rules, with AI models that automate much of the quality checking.

The company said it will use the money to add capacity, push into enterprise and industry work, and fund research. Insight Partners named healthcare, law and software engineering as target industries.

Why traders should care

Snorkel is private, but its numbers are a read on how much the big AI labs are spending on data, not just on chips and power. Reuters pointed to Meta's $14.3 billion investment in Scale AI in June 2025 as the deal that set off this corner of the market, and named Mercor and Surge AI as rivals also drawing investor money. A revenue run-rate that went from about $20 million to more than $350 million in a year, at one supplier, is a sign of how fast that spending has grown. It is also a sign of how concentrated it is: if a few labs cut data budgets, suppliers like this feel it first.

Sources: Reuters via Yahoo Finance, Snorkel AI press release. This is market information, not investment advice.