Snorkel AI reached 3.5 billion dollar valuation after Series E

Snorkel AI Raises 350 Million Dollar Series E

Snorkel AI announced a 350 million dollar Series E round on September 22, 2026, valuing it at 3.5 billion dollars. Co-founder and CEO Alex Ratner disclosed the financing in a company blog post published online early on Tuesday, September 22, 2026.

Insight and S32 led the Series E round, with participation from Third Point, March, Blumberg, Allegis, Standard VC, and Frontline. Existing investors Addition, Lightspeed, Greylock, GV, P7, Wells Fargo, Walden Catalyst Ventures, and Factory also participated in the new financing round.

Co-founder and CEO Alex Ratner said Snorkel's data-as-a-service offering, launched nearly one year earlier, had grown more than 18 times. It crossed the company's 375 million dollar annualized revenue run rate during the same week as the new public funding announcement.

The new Series E follows Snorkel's 100 million dollar Series D funding round, which reached a 1.3 billion dollar valuation when announced earlier on May 29, 2025.

Data 2.0 Strategy and Agentic Data Platform

Snorkel began as a Stanford University research project roughly a decade ago, focusing on the thesis that AI progress would become increasingly data-centric over the long term.

Co-founder and CEO Alex Ratner said data development should be studied as a primary research and technology problem instead of merely a simple staffing and crowdsourcing exercise.

CEO Alex Ratner described Data 1.0 as driven by volumes of simpler data and largely a staffing and logistics problem. In contrast, he described Data 2.0 as driven by quality of more complex data and primarily a research and technology problem.

Snorkel said coding data must approximate complex software problems even a senior engineer might struggle with for days or weeks. Its internal Agentic Data Platform supports many human experts with specialized AI models and individual agents sometimes hundreds per task type.

The company will significantly expand its Open Benchmarks Grants program with a new three million dollar commitment to fund open-source public research datasets, benchmarks, and evaluation artifacts.

Benchmarks Grants Hiring and Future Research Plans

CEO Alex Ratner said specialized agents running quality control alongside expert review accelerate quality control efficiency by more than 50 percent. They improve review accuracy by more than 15 percentage points compared with human reviewers using only off-the-shelf large language models.

Scaled human feedback as weak supervision yields a current two times or greater accuracy improvement over a non-specialized frontier model baseline. Snorkel's Open Benchmarks Grants program will fund open-source datasets, benchmarks, and evaluation artifacts through rolling applications reviewed by external experts.

Listed benchmarks include Terminal-Bench, OSWorld 2.0 with University of Hong Kong's XLANG Lab, and Agent's Last Exam from UC Berkeley's RDI. Others include Continual Learning Bench with UC Berkeley's SkyLab, SlopCode Bench with UW-Madison, and Senior SWE-bench with Wisconsin and Princeton.

With the new cash, Snorkel plans to hire more people and ramp up its services for enterprise and government clients. CEO Alex Ratner said much of Snorkel's future research will focus on data and environment development for AI alignment and safety.