Snorkel AI has raised $350 million in a Series E funding round. The deal values the company at $3.5 billion.
The startup helps AI labs and corporations build training datasets and simulated environments. These tools are used to teach and test AI models.
Simulated environments, also called reinforcement learning environments, let AI models practice tasks in a controlled setting. Its customers include AI labs and large corporations.
Insight Partners and S32 led the round. Existing investors also took part, including Addition, Lightspeed, Greylock, GV, and Wells Fargo.
Valuation Nearly Triples in 17 Months
The new valuation is nearly three times what Snorkel was worth 17 months ago. At that time, it raised $100 million in a Series D round at a $1.3 billion valuation.
The jump means investors now value the company at $2.2 billion more than in its last round.
Snorkel is seven years old. It launched commercially in 2019 after four years of research at a Stanford AI lab.
That research was led by co-founder and CEO Alex Ratner and his team.
The company first offered software that automated data labeling. Data labeling is the process of tagging information so AI systems can learn from it.
Last year, Snorkel changed its business model. It now sells completed datasets to customers, a service it calls data-as-a-service.
The company does not work only as a marketplace for human experts. It uses a hybrid approach, combining its own software and models to create synthetic data with help from subject matter experts.
Revenue Grows Eighteenfold
Snorkel says its annualized revenue run rate is now $375 million. A run rate estimates yearly revenue based on current sales.
That figure is 18 times higher than it was 12 months ago. The company links the growth to strong demand from AI labs seeking high-end training data.
Other AI data companies have also reported fast growth. These firms describe themselves as AI data labs, much like Snorkel.
Mercor's gross annualized revenue has reached $2 billion. Handshake passed the $1 billion mark earlier this year.
Micro1 has also grown quickly. TechCrunch reported that the company scaled to $500 million in gross run rate.
Those companies pay about 60% to 70% of their top-line income directly to the specialists who do the work. This means their net annual revenue is much lower than their gross figures.
Snorkel says its numbers are counted differently. Because it sells reinforcement learning environments and complete datasets, not human labor, payments to its experts fall under cost of goods sold.
That means those payments are not part of its headline annualized revenue number, according to the company.
The $350 million Series E was announced on September 22, 2026. The round brings Snorkel's valuation to $3.5 billion as it continues selling datasets and reinforcement learning environments to AI labs and businesses.
