Snorkel AI raises $350 million at $3.5 billion valuation

Snorkel AI said Sep 22 it raised a $350 million Series E at a $3.5 billion valuation, led by Insight and S32, with participation from Third Point, March, Blumberg, Allegis, Standard VC, Frontline, and existing investors including Addition, Lightspeed, Greylock, GV, P7, Wells Fargo, Walden Catalyst Ventures, and Factory, according to CEO Alex Ratner’s company post.

What the company says the money buys: Snorkel pitches itself as a frontier lab for AI data — data-as-a-service and an “agentic data factory” that builds complex training datasets, evaluation sets, and reinforcement-learning environments for frontier labs, hyperscalers, enterprises, and U.S. government agencies. Ratner wrote that since launching the data-as-a-service offering nearly a year ago, Snorkel has grown over 18x and this week crossed an annualized revenue run rate of $375 million.

Why it matters for markets: Late-stage capital is pouring into the layer that feeds model training and eval — not the models themselves. A $3.5 billion mark on a data-as-a-service stack sits beside neocloud and chip capex as another way private markets price AI infrastructure. That is round size and run-rate, not a tip.

Stake: Frontier labs buying RL environments and expert data; competing data-labeling and eval vendors; Insight/S32 and the prior-round syndicate; compute suppliers whose customers need harder last-mile datasets.

What’s next: How the $375M ARR converts to disclosed GAAP revenue if Snorkel ever files; Open Benchmarks Grants expansion; any customer concentration disclosures.

— Nico Valdez