Snorkel AI has raised $350 million in Series E funding at a $3.5 billion valuation, CEO Alex Ratner told Reuters in an exclusive reported Tuesday, September 22, 2026, as frontier labs push for more complex training data and simulated environments. Insight Partners and S32 led the round, with Addition, Greylock, and Wells Fargo also participating, the company said. The deal values the San Francisco startup at nearly three times the $1.3 billion mark it reached when it raised $100 million in May 2025.
Annualized revenue run-rate has crossed about $350 million, up from roughly $20 million a year earlier, driven by a data-as-a-service line Snorkel launched in September 2025, Reuters reported. Founded in 2019 as a Stanford AI lab spinout, the firm initially sold software and has shifted toward finished datasets and reinforcement-learning environments delivered straight to customers.
Snorkel describes an “agentic data development platform” that pairs human experts with specialized AI models and agents to design, generate, and vet data. Specialists span coding, law, and medicine; coding data is among the largest demand areas. The company says it works with frontier labs, hyperscalers, enterprises, and the U.S. federal government. Ratner told Reuters that useful lab data will still need human input for the foreseeable future, while synthetic and automated methods are required to keep pace with complexity.
The raise lands in a market remade after Meta’s $14.3 billion purchase of a 49% stake in Scale AI in June 2025, with rivals including Mercor and Surge AI also drawing investor interest. Fresh capital is earmarked for researchers and engineers, larger enterprise and government operations, third-party model evaluations, and new verticals and data modalities. Snorkel expects to reach profitability this year.
“Data is becoming more rare, more specialized, more difficult to find,” S32 partner Andy Harrison, who co-led the round, told Reuters. “If you want to train the most frontier, complex and capable models, now you need superior data.”