Micro1’s gross run rate hits $500 million as AI data demand surges

AI data startup Micro1 has expanded its gross annual run rate from $100 million to $500 million over the past eight months, according to a person familiar with the company. The surge is attributed to near-bottomless demand for unique training data from top laboratories and corporations. While the figures have not been officially confirmed by the startup, which did not respond to a request for comment, the growth trajectory highlights the accelerating importance of data in the current AI landscape.

The company retains approximately 60% to 70% of its gross revenue, resulting in a net annual run rate between $150 million and $200 million. Micro1, which began as an AI recruiting business before pivoting to data labelling, competes with larger rivals such as Mercor and Handshake. Mercor reached $2 billion in gross annualised revenue this summer, while Handshake hit $1 billion earlier this year. Despite lagging behind these competitors in total revenue, Micro1’s growth indicates that the market can support multiple players supplying AI training data.

To improve its financial position, Micro1 is increasingly generating synthetic data without human involvement, such as automated descriptions of video content. The startup also sells “off-the-shelf” datasets to multiple clients, a strategy that drives gross margins for this category of data as high as 80% to 90%. This approach allows the company to expand its margins over time, a trend that aligns with researcher hypotheses that future AI spending on data could rival spending on compute.

Founder Ali Ansari has distinguished Micro1 from some of its rivals by stating that the company does not sell data to Chinese model makers. In a recent post on X, Ansari argued that it is “shameful to claim American AI dominance desires while selling millions worth of data to countries that we are in adversarial competition with.” This stance contrasts with critics who argue that distributing off-the-shelf data to Chinese AI developers helps make their models as powerful as top US models.

Micro1 continues to build its data capabilities by hiring domain experts, such as doctors, lawyers, and scientists, on a contract basis. The company is also developing a robotics pre-training dataset by having hundreds of generalists record everyday object interactions in their homes. Additionally, Micro1 uses its experts to evaluate model outputs, a concept known as reinforcement learning gyms, further diversifying its data offerings.

The startup raised its Series A at a $500 million valuation in September last year. TechCrunch understands that Micro1 may have recently secured additional funding at a significantly higher valuation, though this has not been confirmed by the company. As contract sizes grow at an accelerating pace, Micro1 appears well-positioned to benefit from the ongoing boom in AI training data.