Snorkel AI has raised $350 million at a $3.5 billion valuation. Insight Partners and S32 led the round. New investors are March Capital, Blumberg Capital, Allegis Capital, Frontline, Standard and Third Point Ventures. Existing backers include Addition, Greylock, Lightspeed, GV, Factory, Prosperity7, Walden Catalyst and Wells Fargo.
From labeling to “Data 2.0”
Snorkel came out of the Stanford AI Lab in 2019. Its founders helped create the field of data-centric AI, the idea that a model’s quality depends more on its training data than on its architecture. The team’s research runs to more than 250 peer-reviewed papers, cited over 25,000 times.
The company now runs what it calls an agentic data factory. It supplies expert datasets, training environments and benchmarks for frontier AI systems. The focus has moved away from simple labeling. Frontier labs now want complex, research-grade tasks written by experts in medicine, law and software engineering, plus environments where agents can practice multi-step work. Snorkel calls this “Data 2.0.” It launched its expert Data-as-a-Service offering in September 2025.
“We are the frontier lab for agentic data, combining human excellence with over a decade of research and technology,” says co-founder and CEO Alex Ratner.
A crowded, lucrative market
Frontier labs are spending heavily on expert data, and several companies are competing for those contracts. Snorkel’s pitch is research depth plus automation, so experts’ time goes into the hard parts of each task. Its customers are frontier AI labs and enterprises in healthcare, legal services and software engineering.
The San Francisco company didn’t disclose revenue. A $350 million raise shows how much investors think expert data is worth as models run out of easy data from the public web.