Key facts
- AfterQuery, an AI training data startup, has achieved a $3.2 billion valuation.
- This valuation makes it the fastest company in Y Combinator's history to reach unicorn status.
- The company pays experts like doctors, lawyers, and financial analysts to generate specialized AI training data.
- This data focuses on expert judgment and reasoning, which is increasingly scarce.
- Nvidia has used AfterQuery's data to train its Nemotron models.
- Founders Spencer Mateega and Carlos Georgescu are 23 and 22 years old, respectively.
AfterQuery, an AI startup focused on generating expert judgment data, has achieved a valuation of $3.2 billion, making it the fastest company in Y Combinator's history to reach unicorn status. The company, founded 18 months ago by Spencer Mateega and Carlos Georgescu, pays professionals such as doctors, lawyers, and financial analysts to create step-by-step reasoning data that teaches AI models how to make complex decisions.
This specialized data is in high demand as leading AI labs have exhausted readily available text on the open internet and synthetic data has limitations. Nvidia has utilized AfterQuery's data to train its open-source Nemotron models, and clients include Mira Murati's Thinking Machines Lab and legal AI firm Legora. Mateega reported that the company's annual recurring revenue had grown to the hundreds of millions by July 2026, up from $100 million in April.
The company's rapid growth contrasts with competitors like Mercor, which uses an AI interviewer to staff its contractor pool. AfterQuery's competitive edge lies in its custom software that screens submissions for a specific difficulty level and its practice of training its own models on the data before selling it to demonstrate its effectiveness. Rival Mercor is reportedly in talks with Nvidia for a funding round that would value it at $20 billion.