From Zero to $3.2 Billion: How AfterQuery Became Y Combinator’s Fastest Unicorn
In the hyper-competitive world of artificial intelligence, where yesterday’s record-breaking valuation is merely today’s baseline, two young founders have shattered expectations. AfterQuery, an AI data firm founded just 18 months ago by high school friends Carlos Georgescu, 22, and Spencer Mateega, 23, has reached a staggering $3.2 billion valuation, according to sources familiar with the matter.
The milestone marks an incredible ascent. Having joined the Y Combinator startup incubator less than two years ago with no product and no clear roadmap, the duo has seen their valuation jump more than tenfold in just five months. According to YC partner Gustaf Alströmer, this makes AfterQuery the fastest startup in the incubator’s history to reach unicorn status.
The Shift Toward Human Intelligence
AfterQuery’s rapid rise highlights a critical realization among frontier AI labs: as the supply of high-quality public internet data dries up, the “holy grail” for model training has become complex, expert-vetted reasoning data.
While synthetic data offers some relief, it has inherent limitations. Labs are now desperate for domain-specific insights—the nuanced, professional judgment calls made by software engineers, legal experts, and financial analysts. AfterQuery has positioned itself as the gold standard for this human-in-the-loop intelligence.
The company’s impact is already visible at the highest levels of the industry. Its data was utilized in the development of Nvidia’s Nemotron models, and the firm has partnered with industry heavyweights, including former OpenAI CTO Mira Murati’s “Thinking Machines Lab” and legal tech startup Legora. Notably, while some competitors have restricted their reach, AfterQuery also works with major Chinese AI labs.
From Financial Agents to Data Powerhouse
The founders’ path to this valuation was born from a failed experiment. Originally founded in February 2025, AfterQuery was intended to build AI agents for the finance sector. However, while testing their prototypes, Mateega and Georgescu discovered that even the most advanced models struggled with nuanced “white-collar” workflows. They realized the problem wasn’t the architecture, but a lack of expert-level training data capable of mimicking professional decision-making.
They pivoted, shifting their focus from building end-user applications to providing the “picks and shovels” of the AI gold rush.
Quality Control as a Competitive Edge
In a space crowded with data-labeling firms—such as Scale AI and the $20 billion-valued Mercor—AfterQuery distinguishes itself through rigorous quality assurance. Mateega has previously emphasized that while peers often rely on vast pools of contractors, AfterQuery utilizes custom software systems to validate human-generated data.
To ensure their data hits the “Goldilocks” zone—challenging state-of-the-art models without overwhelming them—AfterQuery runs internal research pipelines. By training models on their own data and objectively measuring benchmark performance improvements before handing materials to clients, they provide the empirical proof that labs require before investing millions into training cycles.
As the company continues to scale—boasting recurring revenue in the hundreds of millions as of this summer—it stands as a testament to the new reality of Silicon Valley: in the race to build the smartest machine, the most valuable commodity is human expertise.
AfterQuery declined to comment on the valuation.
