❯ Snorkel AI raises $350 million at a $3.5 billion valuation

Business shiftFounded by Stanford AI Lab researchers in 2019, Snorkel initially sold data-development software. It now increasingly sells finished datasets and reinforcement-learning environments to frontier labs. Domain experts design tasks, scenarios and grading rubrics while specialized models and agents help produce and check the material. Customers pay for training and evaluation outputs, rather than simply buying expert hours.

How it got hereFounded out of the Stanford AI Lab in 2019, Snorkel first sold software for developing training data. After launching data as a service in September 2025, it increasingly began delivering finished datasets and reinforcement-learning environments to model labs. Reuters says its previous $100 million financing in May 2025 valued it at $1.3 billion; demand following the shift helped lift its annualized revenue run-rate.

The financingReuters reported that Snorkel AI raised $350 million in Series E funding, co-led by Insight Partners and S32, at a $3.5 billion valuation. The valuation differs from the $3.2 billion in the supplied lead. Reuters quoted its CEO as saying annualized revenue run-rate exceeded $350 million; that is not completed full-year revenue.

Industry impactThe round will fund research and engineering hires, enterprise and government expansion, and third-party model evaluations. Buying expert-designed data for complex tasks can spare frontier labs from assembling every expert and quality-control process themselves. Suppliers increasingly compete on task design, consistent quality and delivery efficiency, not just the size of their labeling workforce. The reported annualized run-rate is not full-year realized revenue.

▪ SIGNALThe value in AI training data is shifting from labeling volume to expert-designed tasks and verifiable training value.