Snorkel AI Raises $350M to Build Data Infrastructure for Frontier Models

Snorkel AI secures major funding round

Snorkel AI announced it has raised $350 million at a $3.5 billion valuation in a funding round co-led by Insight Partners and S32, with participation from Addition and a roster of new and returning investors. The September 2026 raise will fund expansion of what the company calls its "agentic data factory," which produces specialized datasets and evaluation environments for advanced AI systems.

The San Francisco-based company, which spun out of Stanford's AI Lab in 2019, has positioned itself as a critical supplier to organizations building frontier models. According to the announcement first reported by PRNewswire, Snorkel works with leading AI labs and enterprises on data used for both model training and evaluation.

Why it matters

The funding reflects a fundamental shift in how AI systems are built. As models grow more capable, the bottleneck has moved from simple data labeling to creating complex evaluation scenarios and expert-level training tasks that can take qualified humans hours or days to design properly. Companies that can deliver this "Data 2.0" at scale are becoming essential infrastructure providers in the AI supply chain, particularly as labs race to build increasingly sophisticated agentic systems.

From labeling to expert task design

Snorkel frames the investment around what it describes as a phase change in AI data requirements. The company contrasts today's needs with what it calls the "Data 1.0 era," when building AI primarily meant straightforward labeling work that could be solved by adding more human annotators.

Frontier and agentic systems now require what Snorkel terms "expert agentic tasks, environments, and rubrics" that demand research-level expertise to construct. The company argues that designing these components well has become research work in itself, where quality and complexity determine value rather than volume alone.

Snorkel launched its expert Data-as-a-Service offering in September 2025 and reports rapid growth since then. The company's founding team contributed to pioneering work in data-centric AI, with research spanning more than 250 peer-reviewed papers cited over 25,000 times, according to the announcement.

Investor perspective on data infrastructure

"Snorkel's research-grade approach to AI data, environments, and measurement is becoming an increasingly important ingredient in building capable and reliable AI systems," said Lonne Jaffe, Managing Director at Insight Partners, noting the company's demonstrated growth trajectory.

Andy Harrison, CEO and General Partner of S32, highlighted what he described as a "unique flywheel between human expertise and AI" created by Snorkel's expert-agentic environments.

Expansion plans

The company plans to deploy the capital across several areas: expanding capacity of its data factory operations, accelerating investment in vertical and enterprise AI applications, and extending its core research into new domains and modalities. Snorkel also indicated it will deepen investment in open research initiatives, including its Open Benchmarks Grants program.

The funding round included participation from March Capital, Blumberg Capital, Allegis Capital, Frontline, Standard, and Third Point Ventures as new investors, alongside existing backers Greylock, Lightspeed, GV, Factory, Prosperity7, Walden Catalyst, and Wells Fargo.

Details of the funding were first reported by PRNewswire.