Snorkel AI Triples Valuation to $3.5B as Demand for Complex AI Training Data Surges

Snorkel AI Secures $350M, Valuation Reaches $3.5 Billion

Snorkel AI has closed a $350 million Series E funding round at a $3.5 billion valuation, nearly tripling the $1.3 billion valuation it achieved just 17 months earlier when it raised $100 million in May 20251

1

2

2

. Insight Partners and S32 led the round, with participation from existing investors including Addition, Lightspeed, Greylock, GV, and Wells Fargo1

1

. The San Francisco-based startup, founded in 2019 by researchers from the Stanford AI Lab, has positioned itself at the center of surging demand for complex AI training data as frontier AI labs race to build increasingly capable systems2

2

3

3

.

Revenue Growth Reflects Market Transformation

Snorkel AI's annualized revenue run-rate has crossed $375 million, representing an 18-fold increase over the past 12 months1

1

3

3

. This explosive growth stems from the company's data-as-a-service model launched in September 2025, which shifted Snorkel from selling software for data labeling automation to delivering completed datasets and reinforcement learning environments directly to customers1

1

2

2

. CEO Alex Ratner told Reuters the company had grown from roughly $20 million in annualized revenue a year earlier2

2

. The company expects to reach profitability this year while maintaining its growth trajectory2

2

.

Agentic Data Development Platform Powers Hybrid Approach

Snorkel AI now operates what it calls an agentic data development platform that combines synthetic data and human expertise through a hybrid approach2

2

. The platform pairs tens of thousands of human specialists across fields like coding, law, and medicine with thousands of specialized AI models and agents to create and vet data2

2

. Human experts devise design scenarios, tasks, and grading rubrics, while AI automates much of the labor-intensive quality assurance process2

2

. Ratner emphasized that "100% of the data that labs will get value out of will have some human input in the foreseeable future," but added that "100% of that data will have to use synthetic and automated approaches to keep up with this complexity"2

2

.

Reinforcement Learning Environments Drive Differentiation

Snorkel AI expanded its focus beyond supervised learning to reinforcement learning environments, a more complex AI training approach3

3

. While supervised learning uses datasets containing prompts and correct answers, reinforcement learning datasets contain unanswered questions that AI models must solve independently before receiving feedback3

3

. The company develops AI evaluation rubrics spanning multiple pages that cover requirements like cybersecurity and performance standards for AI-generated code3

3

. Snorkel AI also provides training sandboxes—specialized virtual environments where AI models can be trained—alongside its datasets and evaluation rubrics3

3

.

Market Context and Competitive Landscape

The funding round reflects broader investor enthusiasm for AI training data providers following Meta's $14.3 billion purchase of a 49% stake in Scale AI in June 20252

2

. Other data companies have experienced similar growth trajectories, with Mercor's gross annualized revenue climbing to $2 billion, Handshake hitting $1 billion earlier this year, and Micro1 scaling to $500 million1

1

. However, Snorkel's business model differs from pure labor marketplaces. Since the company sells complete datasets rather than human labor, payments to human experts are accounted for in cost of goods sold rather than reducing headline revenue figures1

1

.

Strategic Investment Plans and Future Direction

Snorkel AI will deploy the fresh capital to hire researchers and engineers while expanding enterprise and government operations2

2

3

3

. The company plans to support third-party AI model evaluations, push into new industry verticals and data modalities, and invest in AI safety initiatives2

2

3

3

. Snorkel also aims to support open-source model evaluation benchmarks3

3

. The company currently serves frontier AI labs, hyperscalers, enterprises, and the US federal government, with coding data representing one of its biggest areas of demand2

2

. Andy Harrison, partner at S32 who co-led the funding, noted that "data is becoming more rare, more specialized, more difficult to find," adding that "if you want to train the most frontier, complex and capable models, now you need superior data"2

2

.