Snorkel AI has raised a $350 million Series E at a $3.5 billion valuation, TechCrunch reports. Insight Partners and S32 led the financing, with Addition, Lightspeed, Greylock, GV and Wells Fargo among the existing investors that participated.
The round values the seven-year-old company at nearly three times the $1.3 billion valuation it received in a $100 million Series D roughly 17 months earlier. The company began as a provider of software for data-labeling automation, following research by co-founder and CEO Alex Ratner and colleagues at Stanford. It launched commercially in 2019.
From labeling tools to delivered data
Snorkel’s current pitch is broader than a labeling tool. The company now provides completed training datasets and simulated environments, an offering it calls data-as-a-service. Its approach uses software and models to generate synthetic data alongside subject-matter experts, rather than operating only as a marketplace for human labelers.
That positioning matters because developers of advanced AI systems increasingly need data and environments tailored to particular tasks, not simply large volumes of generic text or images. Snorkel is selling into that demand from both AI labs and enterprise customers.
Growth figures come from the company
Snorkel says its annualized revenue run rate is now $375 million, up eighteenfold over the past 12 months. The figure is the company’s own measure, and comparisons between data companies can be slippery: TechCrunch notes that some businesses in the category pay a substantial share of headline revenue to the specialists doing the work. Snorkel says payments to its experts are counted as cost of goods sold because it sells datasets and reinforcement-learning environments rather than their labor directly.
The financing gives Snorkel more capital as the market for specialized AI-training inputs becomes more crowded. Its valuation jump shows investors are placing a sizeable bet on companies that can turn domain expertise, synthetic data and evaluation environments into a repeatable product.
Comments