Snorkel AI Valuation Hits $3.5 Billion After $350M

Snorkel AI has raised $350 million in fresh funding at a valuation of $3.5 billion, as demand grows for specialised training data used to develop advanced Artificial Intelligence systems. Chief Executive Alex Ratner told Reuters that the San Francisco-based startup’s annualised revenue run-rate has climbed to approximately $350 million.

The funding round was led by Insight Partners and S32, with participation from existing investors Addition, Greylock and Wells Fargo. The new valuation is nearly three times the $1.3 billion valuation Snorkel AI achieved in May 2025, when it raised $100 million.

The company’s rapid growth reflects a shift in AI development: leading AI labs increasingly need complex, carefully evaluated datasets and simulated environments to train and test more capable models.

Snorkel AI Raises $350 Million at $3.5 Billion Valuation

Snorkel AI’s latest funding round brings in $350 million in new capital and values the company at $3.5 billion, according to Ratner. Insight Partners and S32 led the investment, while existing backers Addition, Greylock and Wells Fargo also participated.

The valuation marks a substantial increase from the $1.3 billion level reported in May 2025. The latest financing comes as investors continue to support businesses supplying data and specialised services to frontier AI developers.

Snorkel plans to use the capital to hire researchers and engineers, expand its enterprise and government operations, support third-party AI model evaluations and enter additional industry sectors and data formats.

Annualised Revenue Run-Rate Reaches About $350 Million

Snorkel AI said its annualised revenue run-rate has exceeded $350 million, compared with roughly $20 million a year earlier. The company attributed the increase to its data-as-a-service business, which it launched in September 2025.

A revenue run-rate is an annualised estimate based on a company’s revenue performance over a particular period. It is not necessarily the same as revenue already recorded over a full financial year.

The reported growth highlights how demand for specialised AI data has become a significant commercial opportunity. However, the company’s expectation of reaching profitability this year remains a forward-looking target rather than a confirmed result.

Why Frontier AI Labs Need More Complex Training Data

AI developers have moved beyond relying only on relatively simple data-labeling tasks. As models become more capable, companies need more specialised examples and challenging environments to train and evaluate their systems.

These requirements can include difficult coding problems, domain-specific legal or medical material, complex tasks and carefully designed evaluation criteria. The data must be useful for training while also meeting quality standards appropriate to the intended application.

Snorkel CEO Alex Ratner said demand has grown as AI developers seek harder, higher-stakes data for increasingly capable systems. The company is positioning its services around this need for specialised data rather than basic annotation alone.

How Snorkel AI’s Data Development Platform Works

Snorkel describes its offering as an “agentic data development platform.” The system combines human specialists with thousands of specialised AI models and agents to create and review datasets.

According to Ratner, human experts design scenarios, tasks and grading rubrics. AI tools then automate much of the labour-intensive quality assurance work, helping produce and check data at greater scale.

The company also works with AI labs to design methods for acquiring high-quality datasets suited to their development needs.

The approach combines human expertise with automated processes. Human specialists contribute domain knowledge and judgement, while AI-based systems assist with generating, processing and checking data.

Human Expertise and Synthetic Data Work Together

Snorkel relies on a network of tens of thousands of specialists across fields including coding, law and medicine. The company says it sells the data products these experts help generate rather than charging customers directly for human labour.

Ratner said this model allows the company to compensate specialists more generously while maintaining margins. The company’s approach reflects a broader debate about how advanced AI systems can obtain enough high-quality data without relying entirely on manual processes.

Ratner argued that human input will remain valuable for the foreseeable future, while synthetic and automated methods will be needed to keep pace with the complexity and volume of data required.

In practice, the balance between human-created, human-reviewed and synthetic data may vary depending on the task, the model being developed and the quality requirements involved.

Coding Data Is a Major Area of Demand

Coding is one of Snorkel AI’s biggest areas of demand, according to the company. Programming-related datasets can involve complex tasks and evaluation criteria that go beyond simple classification or labeling.

Such data can be used to train and assess AI systems designed to assist with software development. As developers seek systems that can handle more involved coding work, the demand for specialised examples and reliable evaluation methods has increased.

Snorkel also serves frontier AI labs, hyperscalers, enterprises and the US federal government, giving it customers across several parts of the AI ecosystem.

AI Training Data Market Attracts Investor Attention

Snorkel’s funding comes amid continued investment in companies that provide human-annotated and specialised training data to advanced AI developers.

The market received heightened attention after Meta purchased a 49% stake in Scale AI for $14.3 billion in June 2025. Other companies, including Mercor and Surge AI, have also attracted investor interest amid strong revenue growth.

The activity reflects the importance of data in developing and evaluating AI systems. As models tackle more complicated tasks, the demand for specialised expertise, carefully constructed datasets and testing environments has become more prominent.

However, companies in this sector can differ significantly in their business models, customer bases, data-generation methods and ability to turn rapid revenue growth into sustainable profitability.

How Snorkel Plans to Use Its New Funding

Snorkel said it will use the new capital to expand its technical and commercial capabilities.

• Hiring: Recruit additional researchers and engineers.
• Enterprise growth: Expand operations serving business customers.
• Government work: Increase its government-related operations.
• AI evaluation: Support third-party evaluations of AI models.
• New markets: Enter additional industry sectors and data modalities.

These plans indicate that Snorkel aims to broaden its role beyond supplying datasets, with a focus on the development and evaluation needs of organisations building or deploying AI.

Snorkel AI Expects to Reach Profitability This Year

Despite prioritising growth, Snorkel AI expects to reach profitability in 2026, according to the company. That goal comes as it plans to expand its workforce and operations using the new funding.

The company’s ability to meet the target will depend on factors such as revenue growth, operating costs, hiring and the economics of producing specialised data at scale. The reported revenue run-rate is a measure of recent business momentum, but it does not by itself establish whether the company is profitable.

For investors, the next stage will involve assessing whether Snorkel can sustain demand while managing the costs associated with expert networks, AI systems and quality assurance.

What Snorkel AI’s Funding Means for the AI Industry

Snorkel AI’s fundraising illustrates how the AI Industry is expanding beyond model development and computing infrastructure into specialised data production and evaluation.

Advanced AI systems require training material that reflects increasingly complex tasks. Companies supplying that material are building services that combine human knowledge, synthetic data and automated workflows.

Snorkel’s strategy centres on coordinating those elements through a platform and selling finished data products to customers. Its new funding will support expansion into additional customer groups and use cases, while the company works toward its stated profitability goal.

The broader market remains competitive, with several companies seeking to serve AI labs and enterprises. The long-term prospects of individual providers will depend on the quality and usefulness of their data, their ability to scale and the commercial value they deliver to customers.

Key Takeaways

• Snorkel AI raised $350 million at a $3.5 billion valuation in a round led by Insight Partners and S32.
• The company said its annualised revenue run-rate exceeded $350 million, up from roughly $20 million a year earlier.
• Snorkel combines human specialists with AI models and agents to create and assess complex training datasets.
• The startup plans to expand hiring, enterprise and government operations, AI evaluations and industry coverage, while targeting profitability in 2026.

Frequently Asked Questions

What is Snorkel AI’s new valuation?

Snorkel AI was valued at $3.5 billion after raising $350 million in fresh funding, according to CEO Alex Ratner.

Who led Snorkel AI’s latest funding round?

The round was led by Insight Partners and S32, with participation from Addition, Greylock and Wells Fargo.

How much is Snorkel AI’s annualised revenue run-rate?

The company said its annualised revenue run-rate has crossed approximately $350 million, compared with around $20 million a year earlier.

What does Snorkel AI do?

Snorkel develops specialised training data and reinforcement-learning environments for AI developers, using a combination of human expertise and automated AI processes.

Why is demand for complex AI training data increasing?

Frontier AI developers need more specialised and challenging data to train and evaluate increasingly capable models, including systems designed for complex coding and other demanding tasks.

How does Snorkel combine human experts and AI?

Human specialists design tasks, scenarios and grading criteria, while AI models and agents help generate and check data and automate quality assurance.

What will Snorkel AI do with the new funding?

The company plans to hire researchers and engineers, expand enterprise and government operations, support third-party model evaluations and enter new industries and data modalities.

Does Snorkel AI expect to become profitable in 2026?

Yes. The company said it expects to reach profitability this year, although that remains a target rather than a confirmed financial result.