Snorkel AI raises $350 million at a $3.5 billion valuation

The round and the revaluation

Insight Partners and S32 led the round. The financing values Snorkel at nearly three times the $1.3 billion valuation it received in its $100 million Series D round.

Existing investors also participated:

The valuation increase is striking, but the more useful number is the relationship between funding and revenue: Snorkel says its annualized revenue has reached $375 million, while the company raised $350 million in the new round.

From labeling software to finished data

Snorkel began with software for automating data labeling. Last year, it shifted toward finished datasets, a business it describes as data as a service.

Its model combines software, internal models and specialists with expertise in specific fields. The software generates synthetic data while specialists work on domain-specific tasks, rather than leaving the company as a marketplace for recruiting experts.

That shift puts Snorkel closer to the companies building the datasets themselves. Several peers have reported rapid growth:

Those figures need a qualification. These companies directly pay the specialists doing the work roughly 60–70% of their revenue, so their actual net annual revenue is substantially lower than the gross figures they report.

Snorkel sells reinforcement-learning environments and finished datasets rather than specialist labor. The company says payments to experts are recorded in the cost of goods and services, not deducted from the annualized revenue figure used in public valuations.

The number behind the headline

My read is that Snorkel’s valuation is a bet on a more durable position in the AI supply chain. Labeling tools can become infrastructure, but finished datasets are closer to the output AI labs are actively buying.

The comparison with Mercor, Handshake and Micro1 is therefore useful but imperfect. Their gross revenue includes large payments that pass through to specialists; Snorkel presents its revenue differently. The announcement is quiet about the number I would want next: how much of Snorkel’s $375 million remains after the expert costs included in its cost of goods and services.

Snorkel began commercial operations in 2019, after four years of research by co-founder and CEO Alex Ratner and his team at Stanford University’s AI lab. The company’s next test is whether its accounting distinction can translate into margins that justify a valuation rising from $1.3 billion to $3.5 billion in 17 months.