Day: September 24, 2026

Snorkel AI has raised $350 million in Series E financing at a $3.5 billion valuation as demand increases for specialized data used to train and evaluate advanced artificial intelligence systems. The financing also highlights the company’s shift from data-labeling software toward completed datasets, evaluations and reinforcement-learning environments for AI developers.

Key Takeaways

• Snorkel AI announced a $350 million Series E financing on September 22, 2026, at a $3.5 billion valuation.
• Insight Partners and S32 led the financing, with new and existing backers participating.
• The company said its annualized revenue run rate crossed $375 million.
• Snorkel AI has expanded from data-development software into specialized datasets and reinforcement-learning environments.
• The company plans to expand its research, engineering, enterprise and government operations.

Snorkel AI announced the new financing on September 22, placing its valuation at nearly three times the $1.3 billion figure attached to its previous funding round in 2025.

Insight Partners and S32 led the Series E. Other participants included Third Point, March, Blumberg, Allegis, Standard VC and Frontline, along with existing backers Addition, Lightspeed, Greylock, GV, P7, Wells Fargo, Walden Catalyst Ventures and Factory, according to the company.

The size of the round comes as developers of advanced AI systems seek increasingly specialized data for model training, evaluation and reinforcement learning. Those requirements extend beyond the large collections of general-purpose text, images and code used during earlier stages of model development.

Snorkel AI Reports a $375 Million Annualized Run Rate

One of the most closely watched figures in Snorkel AI’s announcement was its reported annualized revenue run rate.

The company said the figure crossed $375 million during the week of the financing announcement. Snorkel AI attributed the increase partly to its data-as-a-service offering, which it introduced in September 2025, and said the run rate had grown more than 18-fold over the previous 12 months.

Reuters separately reported annualized revenue above $350 million, compared with about $20 million a year earlier.

The distinction between annualized run rate and recorded annual revenue is significant. A run rate projects recent business performance over a longer period. It does not represent revenue already recognized during a completed 12-month financial period.

Snorkel AI’s expansion comes as companies across the AI training data market compete to provide increasingly technical material for model developers.

The sector includes businesses using combinations of subject-matter experts, synthetic data, software tools and automated systems to create data designed for specific AI tasks.

At the same time, AI infrastructure startups have continued developing technologies that support model training, deployment, computing and data preparation as artificial intelligence systems become more complex.

The Business Has Expanded Beyond Data-Labeling Software

Snorkel AI traces its origins to research conducted at Stanford and launched commercially in 2019.

Its earlier work centered on programmatic data labeling, an approach intended to help organizations create training data without depending entirely on manual annotation.

The company’s commercial model has since broadened.

Snorkel AI now supplies completed datasets and reinforcement-learning environments directly to customers. These products can be used to train and evaluate AI models on tasks that require specialized reasoning, technical expertise or detailed performance criteria.

According to Reuters, the company draws on specialists in fields including software engineering, medicine and law while using AI systems to assist with data creation and quality control.

Human experts can develop tasks, scenarios and grading criteria, while automated systems can support parts of the review and verification process.

That approach reflects changing requirements in AI training.

Conventional annotation projects may involve assigning categories to images, labeling text or identifying specific features in datasets. More advanced AI development can require experts to create realistic problems, evaluate model responses and determine whether an answer is technically correct rather than simply plausible.

Chief Executive Alex Ratner told Reuters that human involvement is expected to remain part of the data-development process even as automation increases.

“Our strong view is that 100% of the data that labs will get value out of will have some human input,” Ratner said.

The company has also described the use of specialized models and AI agents to support quality-control work. Performance figures associated with those internal systems are company-reported and have not been independently established through public financial disclosures.

Complex AI Models Are Changing Data Requirements

The demand for more specialized datasets reflects changes in how increasingly capable AI models are trained and evaluated.

Large-scale pretraining has traditionally relied on substantial volumes of text, images, code and other digital material. Later stages of development can involve reinforcement learning, evaluations and specialized environments designed to measure how well models perform on specific tasks.

Software engineering is one area where these requirements have become more complex.

Coding has become an important part of Snorkel AI’s business as AI developers seek training environments that more closely resemble real software-engineering work.

Rather than relying only on individual programming questions, advanced evaluations can require models to navigate codebases, identify defects, modify software and complete tasks against defined technical criteria.

Snorkel AI has also participated in research involving AI benchmarks and evaluation environments, including work focused on software engineering.

The company says its broader research activities include coding, financial reasoning and AI agent performance.

Snorkel AI has said it serves frontier AI laboratories, cloud providers, enterprises and U.S. government organizations. The company has not publicly disclosed complete financial information for individual customer relationships or contracts.

The growth of specialized evaluation environments also illustrates a broader challenge facing AI developers. As models improve on standard benchmarks, developers need more difficult tests that can distinguish between systems capable of producing convincing responses and those able to complete complex work accurately.

Snorkel AI Plans Further Research and Operational Expansion

Snorkel AI plans to use the new financing to add researchers and engineers while expanding its enterprise and government businesses, according to Reuters.

The company also intends to develop additional types of data, move into more industry-specific applications and support third-party evaluations of AI models.

Ratner told Reuters that Snorkel AI expects to reach profitability by the end of 2026. That remains a company forecast rather than a completed financial result.

The new round follows Snorkel AI’s $100 million Series D in May 2025, when the company was valued at $1.3 billion. Its latest $3.5 billion valuation represents an increase of about 2.7 times in roughly 16 months.

The financing also reflects how Snorkel AI’s business has changed since its early focus on programmatic labeling.

Its current offerings increasingly center on producing completed data products, model evaluations and reinforcement-learning environments rather than solely providing software that customers use to prepare their own training data.

The company’s reported $375 million annualized run rate offers one measure of that transition, although the figure remains an annualization of recent business activity rather than recognized full-year revenue.

With the $350 million Series E, Snorkel AI plans to expand the research, engineering and specialized data operations supporting that business model.

Frequently Asked Questions

How much did Snorkel AI raise?

Snorkel AI announced a $350 million Series E financing on September 22, 2026. Insight Partners and S32 led the round, with a group of new and existing backers participating.

What is Snorkel AI’s valuation?

The financing values Snorkel AI at $3.5 billion. The company’s previous $100 million funding round in May 2025 carried a valuation of $1.3 billion.

How much revenue does Snorkel AI generate?

Snorkel AI said its annualized revenue run rate crossed $375 million in September 2026. The figure annualizes recent business activity and should not be interpreted as recognized revenue for a completed financial year.

What does Snorkel AI do?

Snorkel AI develops specialized datasets, evaluation tools and reinforcement-learning environments used in AI development. Its operations combine software, automated systems and human subject-matter expertise.

How will Snorkel AI use the new financing?

The company plans to expand its research and engineering teams, develop additional data products and grow its enterprise and government operations. It also intends to support more AI model evaluation work across specialized fields.