Snorkel AI has raised a $350 million Series E at a $3.5 billion valuation, nearly triple the valuation from its previous round 17 months ago. Insight Partners and S32 led the financing, with Addition, Lightspeed, Greylock, GV and Wells Fargo also participating.
The seven-year-old company began with software that automated data labeling. Last year it shifted toward delivering completed datasets and simulated reinforcement-learning environments, combining synthetic generation with work from subject-matter experts. That makes Snorkel more like a data supplier than a software toolkit customers must operate themselves.
Snorkel reports a $375 million annualized revenue run rate, 18 times its level a year ago. The number is a company-provided current-rate measure rather than audited annual revenue. Rival data providers often report gross figures before paying 60% to 70% to outside specialists; Snorkel says payments to experts instead appear in its cost of goods sold because it sells finished products rather than labor.
The funding reflects a growing bottleneck for model developers. Once broadly available web data is exhausted, labs need specialized examples, expert judgments and interactive environments to improve reasoning and agent behavior. Snorkel’s new valuation assumes that demand will persist even as models become better at generating some of their own training material.