Snorkel AI raises $350M Series E at $3.5 billion valuation

By Splitfeed EditorialLast updated Sep 26, 2026, 20:18 UTC

On September 22, 2026, Snorkel AI closed a $350 million Series E funding round that set its valuation at $3.5 billion, nearly tripling the company's previous figure. The seven-year-old startup operates a data-as-a-service model, supplying the complex training datasets that AI labs use to build and refine large models. The company framed the round around plans to scale what it calls a frontier data factory, and the raise was accompanied by a reported revenue run rate of $375 million, up 18 times in a single year, according to the Pro-AI reactions circulating after the announcement.

The round arrives at a moment when AI developers are increasingly competing not just on model architecture but on the quality and volume of training data. That shift gives a company like Snorkel AI a distinct position in the supply chain, sitting between raw data and the frontier labs that consume it. What is genuinely in dispute is whether the headline metrics reflect durable business fundamentals. The gap between a run-rate figure and actual booked revenue is a live question, and the speed of the valuation increase, tripling in what appears to be a relatively short period, invites scrutiny about whether investor enthusiasm is tracking real demand or amplifying it.

The Pro-AI camp treats the raise as confirmation that infrastructure and tooling are where the real value in AI sits right now. Accounts including Five O'Clock Tech and SkyPilot framed it as a picks-and-shovels story, with INS AI Solutions pointing to the 18x revenue growth as evidence for teams whose AI deployments have stalled on reliability concerns. The Anti-AI camp reads the situation with more suspicion. Artologica described the language of thesis-driven investment as a cover for corruption-driven contracts, while ROU Megastructure Enthusiast mocked the valuation scale by sardonically announcing a fictional billion-dollar raise of their own. The Middle Ground camp occupies a more analytical position, with spaisee noting that specialized datasets may now be worth more than larger models, and Marcus Schuler flagging the specific concern about whether run-rate metrics translate into booked revenue.

The argument is likely to sharpen as Snorkel AI deploys the new capital and reports subsequent financials. Actual revenue figures against the stated run rate, and whether frontier labs renew or expand data contracts, would go a long way toward settling the dispute between those who see a structural shift in AI value chains and those who see a valuation inflated by momentum.