Snorkel AI has raised $350 million at a $3.5 billion valuation as demand for training data and simulated environments accelerates. The Series E nearly triples the $1.3 billion valuation the company received when it raised $100 million 17 months earlier.
Insight Partners and S32 led the new round. Existing investors Addition, Lightspeed, Greylock, GV and Wells Fargo also participated. The financing gives Snorkel a prominent valuation in a market where AI developers need both specialized human expertise and synthetic material to train and test increasingly capable systems.
Snorkel Moves Beyond Labeling Software
Snorkel began with software for automating data labeling. Last year, it shifted toward delivering completed datasets as a service, rather than asking customers to operate the tooling themselves. Its current offering combines software-generated synthetic data with work from subject-matter experts.
The company also sells reinforcement-learning environments, giving customers simulated settings in which models can practice or be evaluated. That mix distinguishes Snorkel from businesses built primarily as marketplaces that connect AI labs with human specialists.
Snorkel launched commercially in 2019 after four years of research by co-founder and chief executive Alex Ratner and his team at a Stanford AI laboratory. The company’s evolution from research and software into a managed data service reflects where it says customer demand has moved.
The funding announcement does not disclose how the new capital will be allocated. It also does not provide customer counts, contract duration or the balance between dataset work and reinforcement-learning environments. The round confirms investor appetite, while leaving the operating shape of the business only partly visible.
Revenue Growth Comes With Accounting Caveats
Snorkel says its annualized revenue run rate has reached $375 million, an eighteenfold increase over the past 12 months. That is a company-reported pace based on current activity, not audited annual revenue reported after a completed fiscal year.
TechCrunch places the figure beside rapid growth at other AI data companies. Mercor’s gross annualized revenue has reportedly reached $2 billion, Handshake has reached $1 billion and Micro1 has scaled to $500 million. Those headline numbers are not directly comparable with net revenue.
Human-expert marketplaces typically pass roughly 60 to 70 percent of their top-line income to the specialists performing the work, according to the report. Their actual net annual revenue is therefore substantially below the gross figure.
Snorkel says payments to its experts are recorded in cost of goods sold because it delivers datasets and simulated environments rather than selling access to human labor. That accounting treatment makes its headline revenue structurally different from a marketplace’s gross transaction volume. It does not remove the labor cost; it places the expense elsewhere in the financial picture.
The distinction is crucial when valuations are justified with fast-growing annualized figures. Comparing top-line run rates without examining what each company retains can make businesses with different economics appear more similar than they are.
A Higher Valuation Raises the Evidence Bar
The $3.5 billion valuation implies that investors expect demand for high-quality training material to persist. Snorkel’s hybrid approach is designed for customers that need domain expertise while also using software and models to generate data synthetically at scale.
The source does not disclose profitability, cash use, customer concentration or net retention. It also does not show what portion of the eighteenfold growth came from recurring contracts rather than a smaller number of large projects. Those omissions do not negate the reported growth, but they limit what the round reveals about durability.
Snorkel’s rise illustrates how the AI boom creates value beyond model developers and computing providers. Training and evaluation depend on the data, environments and expert judgments surrounding a model. The new financing prices that supporting layer as a major business in its own right.
The next test will be whether the company can convert a rapid annualized run rate into sustained revenue with attractive margins. For now, the Series E shows that investors are willing to pay substantially more for Snorkel’s position in the data pipeline, while the available numbers still require careful comparison with other firms in the category.
