Set a Record with Rapid Value Growth
AfterQuery, which provides datasets and evaluation environments for training AI models, raised its market valuation to 3.2 billion dollars following its latest funding round. The San Francisco-based company had reached a valuation of 300 million dollars in its 30 million dollar Series A round announced in April 2026. This means the company increased its valuation more than 10-fold in less than five months.
According to a statement by Y Combinator partner Gustaf Alströmer, AfterQuery became the startup to reach unicorn status in the shortest time from its founding in the accelerator program’s nearly twenty-year history. The company’s founders, Carlos Georgescu (22) and Spencer Mateega (23), joined Y Combinator’s 2025 Winter batch 18 months ago.
Annual Recurring Revenue Surpassed 100 Million Dollars
AfterQuery announced in April 2026 that its annual recurring revenue (ARR) had exceeded 100 million dollars.
Maintaining its revenue growth rate in the following months, AfterQuery continues to collect data through its network of approximately 100 thousand verified individuals specializing in fields such as finance, medicine, law and software engineering.
The platform differs from traditional data-labeling methods by transferring experts’ thought processes, decision-making stages and reasoning methods to AI systems, and serves leading laboratories in the industry. AfterQuery’s customers include major companies such as OpenAI, Google DeepMind, Microsoft AI, Legora and Motif Technologies. Nvidia, which reached a revenue record exceeding 92 billion dollars in its quarterly financial results, also uses datasets provided by AfterQuery and training environments created for its Nemotron models.
Direct Transfer of Expert Knowledge to Advanced Models
The depletion of publicly available data on the internet has made high-quality and original data a critical element in training next-generation AI models.
Through the expert network it has built, AfterQuery enables AI agents not only to provide accurate answers to questions but also to complete complex professional tasks at an expert level.
The startup provides multimodal datasets in addition to text data, as well as custom simulation environments for reinforcement learning (RL) processes. These capabilities allow AI developers to put their models through performance tests.