Snorkel AI Raises $350M, Valuation Climbs to $3.5 Billion

Snorkel AI raised $350 million on Sept. 22, at a valuation its investors put at $3.5 billion post-money, according to the company's own announcement, in a round co-led by Insight Partners and S32. That is roughly 2.7 times the $1.3 billion valuation Snorkel carried at its Series D about seventeen months earlier, according to TechCrunch, which first reported the prior figure.

The jump in valuation is the number that will travel. The more informative number is the one Snorkel volunteered alongside it: an annualized revenue run rate of $375 million, which the company says is eighteen times what it was a year earlier. Set against the new $3.5 billion valuation, that run rate implies a multiple of about 9.3 times trailing revenue — a figure nobody outside Snorkel has audited, since it comes from the company's own disclosure rather than a filing.

Founded inside Stanford's AI Lab in 2019

Snorkel was spun out of Stanford University's AI Lab in 2019 by a team led by co-founder and chief executive Alex Ratner, whose research group built tools for labeling training data programmatically instead of by hand. "The teams pushing the frontier want a research data partner who pioneers the science of data development," Ratner said in the announcement. "That's what Snorkel was built to be: the frontier lab for agentic data, combining human excellence with over a decade of research and technology."

The business Snorkel is now selling is not the one it started with

Snorkel spent its first several years selling data-labeling software that customers ran themselves. In September 2025 it launched a data-as-a-service product instead, pairing synthetic data generation with paid subject-matter experts to build what the company calls "expert-agentic environments" — training tasks, rubrics and evaluations complex enough that even qualified humans need hours or days to construct one. Roughly 60 to 70 percent of that service's top-line revenue goes to the specialists who build the tasks, according to TechCrunch's reporting on the round.

Snorkel's growth rate against three named rivals

Snorkel is not the only company selling expert-built data to AI labs, and TechCrunch's report on the round put three competitors' revenue on the record for comparison. None of the three figures below have been independently audited; each is the company's own disclosure.

Company	Reported annualized revenue
Mercor	$2 billion (gross)
Handshake	$1 billion (2026 milestone)
Snorkel AI	$375 million
Micro1	$500 million (gross)

1. Series D (~17 months ago)1.3 $B
2. Series E (Sept. 22, 2026)3.5 $B

Source: TechCrunch, accessed 2026-09-27

Who financed the round, and who is new to the cap table

The round included new investors March Capital, Blumberg Capital, Allegis Capital, Frontline, Standard and Third Point Ventures, alongside existing investors Addition, Greylock, Lightspeed, GV, Factory, Prosperity7, Walden Catalyst and Wells Fargo, according to the announcement. Lonne Jaffe, managing director at Insight Partners, said the firm was backing Snorkel's push into "longer-running, complex work in high value industries like healthcare, law, and software engineering." Andy Harrison, chief executive and general partner at S32, said Snorkel's approach "ensures the highest frontier model success rate."

The company said the money will expand the capacity of its data factory, fund investment in vertical and enterprise AI, and extend its research into new domains and modalities, including its Open Benchmarks Grants program. That is a broader mandate than the Series D funded, and it is also vaguer: Snorkel has not said how it will divide the $350 million among those three goals, or when it expects to next report a revenue figure. Recent rounds elsewhere in AI-native enterprise software, like Lightfield's $47 million Series A for an AI-agent CRM, have shown the same pattern of a big number attached to a small set of disclosed metrics. Positron AI's own 2.7-times jump to a $5 billion valuation in seven months is the closest recent comparison for how fast AI infrastructure valuations have moved in 2026.

The next figure worth watching is not the valuation. It is whether Snorkel, or any of the three competitors named alongside it, ever discloses a net margin once the cost of the human specialists is subtracted from the revenue line.