Snorkel AI, which offers an “agentic data development platform” with humans and AI agents to create and vet data, raised $350M at a $3.5B valuation
Context & Ripple Effects
Snorkel AI began as a data-labeling service, reaching a $1 billion valuation in 2021, then repositioned around AI evaluation and human expert judgment in its 2025 evaluation-tools pivot. The new financing prices that broader data-development strategy far above its earlier funding rounds.
The company also entered a competitive opening: in June 2025, Snorkel, Labelbox and Turing reported more client interest amid concerns about Scale AI's independence after Meta's investment, as covered in the scramble among Scale AI rivals.
First-order effects
- Snorkel AI gains $350 million to expand its platform for creating and vetting AI data, strengthening the capital base behind its human-and-agent workflow.
- The $3.5 billion valuation gives Snorkel a stronger market signal as an enterprise provider of AI evaluation and data development rather than a standalone labeling vendor.
Second-order effects
- Labelbox and Turing face a better-funded Snorkel in enterprise accounts already reassessing data suppliers after the Scale AI independence concerns documented in 2025.
- Buyers seeking both training-data creation and model-quality review gain a more heavily financed alternative, raising pressure on suppliers to combine those functions rather than sell labeling alone.
Third-order effects
- If enterprise demand continues to favor human-reviewed AI outputs, data development and evaluation may consolidate into a strategic AI-infrastructure layer, with capital concentrating in platforms that coordinate experts and agents.
- The funding reinforces a shift away from treating labeled datasets as a one-off input toward ongoing systems for measuring and improving model behavior.
The trend: AI data suppliers are evolving from labeling vendors into capital-intensive platforms for continuous data generation, evaluation and human oversight.