Snorkel AI
NativeAI NativePrivateUpdated November 15, 2026The programmatic data labeling company that is upstream of every AI model that matters.
- Website
- snorkel.ai
- Founded
- 2019
- CEO
- Alex Ratner
- HQ
- Redwood City, CA, United States
- Stage
- Series E
- Revenue
- Not disclosed
- Total Funding
- $585M
- Employees
- 201-500
Executive Summary
Updated Nov 12, 2026
What they do
- Builds expert-authored training datasets, benchmarks and evaluations for frontier AI models
- Supplies agentic RL environments, such as sandboxed coding tasks, to train and test agents
- Still offers its legacy programmatic labeling platform Snorkel Flow and the Snorkel Evaluate product
Why it matters
- Says revenue grew 18x to a $375M annualized run rate after its data-as-a-service pivot
- Raised a $350M Series E at a $3.5B valuation in September 2026
- Is independent of model makers, unlike Meta-backed Scale AI, which some frontier labs prefer
Best fit
- Frontier and neolab model teams buying expert post-training data and RL environments
- Enterprises in regulated fields needing domain-specific evaluation and benchmark datasets
- AI teams wanting research-grade agent benchmarks for coding, insurance or clinical workflows
Watch-outs
- Pivot to managed services makes the future of Snorkel Flow software customers unclear
- Headline revenue is annualized run rate, not booked revenue, and enterprise share is undisclosed
- Revenue likely depends heavily on a few frontier labs, none of them publicly named
Products & Offerings
Expert Data-as-a-Service
Managed service that builds custom expert-curated datasets, evaluations and post-training data for frontier labs and enterprises.
Snorkel Data Series
Ready-made, expert-authored datasets and agentic environments, such as SWE-Bench-CLI+, with difficulty tiers, refreshed quarterly.
Snorkel Evaluate
Enterprise evaluation product for building benchmark datasets and specialized evaluators and for finding error patterns in LLM and RAG systems.
Snorkel Flow
Data development platform that labels and manages training data programmatically, built on Stanford weak-supervision research.
Snorkel Benchmarks & Leaderboards
Public frontier benchmarks, including Agentic Coding 2.0, plus contributions to Terminal-Bench, used to measure agent performance on real-world tasks.
Recent Releases
- R4mModelHugging FaceMar 2026
Latest News
Palo Alto Networks · 13 days ago
PositiveKED Global · 21 days ago
PositiveBigGo Finance · 21 days ago
Positiveqz.com · 1 month ago
PositiveYahoo Finance · 1 month ago
PositiveUpdated Nov 12, 2026
Snorkel AI is a credible, well-funded independent supplier of expert training data, RL environments and benchmarks. It competes directly with Scale AI, Surge AI, Mercor, Turing and Handshake, and frontier labs and enterprise teams building specialized agents should run a structured evaluation against those vendors. Buyers who want self-serve programmatic labeling software should confirm Snorkel Flow's roadmap and support commitments first, because the company has repositioned itself as a managed data lab.
Signal Provenance
Latest observation date and source for each signal category
| Signal Category | Last Updated | Source |
|---|---|---|
Funding Signals Last funding round, Funding stage | Nov 12, 2026 | Cited web research |
Hiring Velocity Headcount, Headcount growth | Nov 15, 2026 | Research agents (estimate) |
Technical Output Research papers, GitHub activity | Nov 15, 2026 | Hugging Face Hub + news feed + research agents |
News Sentiment News mentions, News sentiment | Nov 15, 2026 | Verified news feed |
Partnership Signals Enterprise partnerships, Distribution partners | Nov 15, 2026 | Research agents (estimate) |
Revenue Indicators ARR estimate, Revenue growth | Nov 15, 2026 | Research agents (estimate) |
Leadership Signals CEO tenure, Founder-led | Nov 15, 2026 | Research agents (estimate) |
Competitive Position Category rank, Market share | Nov 15, 2026 | Research agents (estimate) |
Why this Bluescore?
6 dimensions, each scored 0–10 against all 250 companies. Each adds its weight × score to the Bluescore — see methodology
Snorkel AI's Bluescore of 52.7 rests most on Talent & Research (7.8, top 14%) and Product Depth (6.3, top 21%). Its weakest area is Business Model & Revenue (3.5, 18th percentile).
In plain English
Snorkel AI's Bluescore of 52.7 is lifted most by Talent & Research, where it scores 7.8 of 10 and sits at the 85th percentile among the AI250, and by Product Depth, at 6.3 of 10 and the 79th percentile. Business Model & Revenue holds it back most, with a score of 3.5 of 10, the 18th percentile, and 13.8 points still left of its 21 weight. Because that dimension has the largest unclaimed headroom of any category, improving Business Model & Revenue would move the overall score more than any other change.
Written by AI from the scores below and checked against them; updated nightly.
Traction Signals
Last Round
Series E
Risks
Updated Nov 12, 2026