Snorkel AI, a company assisting AI laboratories and businesses in creating training datasets and simulated environments, has secured $350 million in a Series E funding round, achieving a valuation of $3.5 billion.

This latest funding round, spearheaded by Insight Partners and S32, values the seven-year-old company at almost three times the $1.3 billion valuation it received after raising $100 million in a Series D round 17 months prior. Current investors, including Addition, Lightspeed, Greylock, GV, and Wells Fargo, also took part in this round.

Initially, Snorkel offered software for automating data-labeling processes, but last year it transitioned to delivering finalized datasets to its customers, branding this service as data-as-a-service. Instead of functioning solely as a marketplace for human expertise, Snorkel employs a combined strategy, employing its software and models to create data synthetically along with input from subject matter experts.

According to Snorkel, its present annualized revenue run rate has reached $375 million, representing an 18-fold increase over the past year. This surge is driven by the unquenchable need for high-quality training data from AI labs.

Other data firms positioning themselves as AI data laboratories are experiencing a similar surge in growth. Mercor has seen its gross annualized revenue grow to $2 billion, Handshake surpassed the $1 billion mark earlier this year, and TechCrunch has reported that Micro1 has expanded to $500 million. Since these companies allocate approximately 60% to 70% of their gross revenue directly to the domain experts performing the tasks, it is crucial to understand that their actual net annual revenue is significantly lower than those reported gross figures.

Given that Snorkel offers reinforcement learning (RL) environments and full datasets instead of human labor, payments to its human experts are incorporated into its cost of goods sold rather than reported annualized revenue figures, as stated by the company.

Snorkel began commercial operations in 2019 after four years of research by co-founder and CEO Alex Ratner and his team at a Stanford AI lab.