From TechCrunch · Marina Temkin · · 1m 30s
Snorkel AI hits $3.5B valuation on data boom
The AI training data startup triples its value in 17 months as companies race to build better models.
In brief
The AI training data startup triples its value in 17 months as companies race to build better models. Snorkel AI's explosive growth reflects how critical high-quality training data has become for building competitive AI systems. Originally reported by TechCrunch.
Snorkel AI triples in value
The startup raised $350 million at a $3.5 billion valuation, nearly triple its $1.3 billion value from 17 months ago. Insight Partners and S32 led the round.
From labeling to complete datasets
Snorkel shifted from software for data labeling to selling finished datasets and simulated environments. It now uses a hybrid approach combining synthetic data generation with human experts.
Revenue exploded eighteenfold
Snorkel's annualized revenue run rate jumped to $375 million over the last 12 months. The growth is driven by AI labs' massive demand for high-quality training data.
Competitors also scaling fast
Other AI data companies are seeing similar explosive growth. Mercor hit $2 billion in gross annualized revenue, Handshake reached $1 billion, and Micro1 scaled to $500 million.
Different accounting for data
Unlike competitors that pay 60-70% of revenue to human workers, Snorkel accounts for expert payments in cost of goods sold. This makes its net revenue substantially higher than headline figures.
Stanford roots, commercial launch
Snorkel launched commercially in 2019 after four years of research by co-founder and CEO Alex Ratner at Stanford's AI lab. The company is now seven years old.
The main idea
Snorkel AI's explosive growth reflects how critical high-quality training data has become for building competitive AI systems.
