TL;DR — Key Takeaways
- Snorkel AI raised $350 million in a Series E round at a $3.5 billion valuation, nearly three times its May 2025 valuation.
- Snorkel says it has grown more than 18-fold since launching its data-as-a-service business about a year ago, with annualized revenue run rate surpassing $375 million.
- The company plans to invest the new funding in open AI benchmarks, specialized AI systems and more complex training and evaluation data, including work focused on AI safety.
Snorkel AI announced it raised $350 million in a Series E round at a $3.5 billion valuation. The company is known for its tools and services for AI developers to create, label and evaluate training data. Over the past year, Snorkel has expanded from primarily selling data development software into delivering finished datasets and evaluation environments to AI labs, enterprises and government customers.
That newer part of Snorkel’s business has grown quickly. Since launching its data-as-a-service business about a year ago, Snorkel said it has grown more than 18-fold and that its annualized revenue run rate surpassed $375 million. Its offerings now include reinforcement learning environments designed for training and evaluating more advanced AI systems.
Snorkel CEO Alex Ratner wrote that the company began as a Stanford research project a decade ago with the thesis that AI progress would become increasingly data-centric, making data development a research and technology problem. The company argues that as AI models become more capable, developers need training and evaluation tasks that are harder to produce than conventional labeled datasets. The company says some coding environments, for example, must resemble software engineering tasks that could take experienced developers days or weeks to complete while also accounting for model errors, reward hacking and other failure modes.
Building these more complex training and evaluation tasks requires expertise in both the subject matter and the ways AI models can fail. Snorkel’s platform combines subject matter experts with specialized AI models and agents to build tasks, environments, grading rubrics and quality controls. Snorkel draws on a network of thousands of specialists in fields including coding, law and medicine, while selling completed data products rather than billing for expert labor.
Insight Partners and S32 led the latest funding round, with participation from Third Point, March and several existing investors. The new valuation is nearly three times the $1.3 billion Snorkel reached when it raised $100 million in May 2025.
Ratner said the new funding will support Snorkel’s research around AI data development, including work on open benchmarks, specialized AI systems and training and evaluation data for AI safety. The company plans to significantly expand its Open Benchmarks Grants program, which provides funding and data development support for independent AI benchmarks. Snorkel also plans to invest in data and environments for specialized models built around particular enterprise workflows and use cases, an area Ratner expects to account for a larger share of AI data development over the next several years.
