1 hour ago
30
AfterQuery, an AI data company that barely existed 20 months ago, just hit a $3.2 billion valuation. That makes it the fastest company in Y Combinator’s nearly two-decade history to reach unicorn status, a milestone that typically takes startups years to achieve, not months.
The company closed a $30 million Series A in April 2026 at a $300 million valuation. By September 2026, it had multiplied that figure by more than ten.
What AfterQuery actually does
The company’s core business is deceptively simple: it provides human-curated training datasets and reinforcement learning environments to AI labs. Large language models are only as good as the data they learn from, and most of the internet’s publicly available text has already been scraped dry.
AfterQuery’s edge is its network of roughly 100,000 verified professionals across finance, medicine, law, and software engineering. These aren’t random crowdsourced workers labeling images. They’re domain experts generating the kind of nuanced, high-quality data that separates a mediocre AI model from one that can actually reason about complex problems.
The client list reads like a who’s who of frontier AI development. OpenAI, Google DeepMind, and Microsoft AI all use AfterQuery’s datasets. The company has also contributed to Nvidia’s Nemotron models and maintains relationships with Chinese AI labs, positioning itself as a Switzerland-like data provider in an increasingly fragmented geopolitical landscape.
Co-founders Spencer Mateega and Carlos Georgescu launched the company in January 2025. They originally built an AI financial agent but pivoted after running into the messy reality that nuanced financial workflows don’t lend themselves to simple automation.
The numbers behind the hype
AfterQuery reported over $100 million in annual recurring revenue back in April 2026. By July, the company described that figure as having climbed to “hundreds of millions.” Revenue growing that fast with total disclosed funding of roughly $30 to $34 million means the company is operating with remarkable capital efficiency.
The 10x valuation jump from $300 million to $3.2 billion in five months is eye-popping, but it’s worth noting that the AI data market has its own peculiar dynamics right now. Every major AI lab is racing to acquire proprietary training data as freely available internet text becomes insufficient for next-generation models. Reddit signed a $60 million annual deal with Google for training data access. AfterQuery is selling something arguably more valuable: expert knowledge that doesn’t exist anywhere on the open web.
Disclosure: This article was edited by Editorial Team. For more information on how we create and review content, see our Editorial Policy.
English (US) ·