AI data labeling startup Micro1 reports $500M gross run rate amid training-data demand surge

Micro1, a four-year-old AI data-labeling startup, 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 company retains roughly 60% to 70% of gross revenue, which would place its net annual run rate between $150 million and $200 million.

The surge in revenue reflects broader industry dynamics, as top AI labs and corporations seek unique training data to improve model performance. Micro1 competes in a cohort of data-labeling startups that employ domain experts such as doctors, lawyers, and scientists on a contract basis to annotate and curate datasets.

While Micro1’s growth is notable, it trails larger rivals in the space. Competitors like Mercor reported $2 billion in gross annualized revenue this summer, and Handshake reached $1 billion earlier in the year. Despite this gap, the demand for AI training data appears sufficient to support multiple players, with some researchers speculating that future AI spending on data could rival spending on compute.

Micro1’s revenue growth is expected to continue, driven by increasing contract sizes and expanding margins. The company is also generating synthetic data without human involvement, such as automated descriptions of video content. Additionally, some datasets are sold to multiple customers as "off-the-shelf" products, achieving gross margins as high as 80% to 90% for this segment, according to a person familiar with the startup’s finances.

The practice of selling the same datasets to multiple clients has drawn criticism, with concerns that such data could be used by Chinese AI developers to enhance their models. Micro1’s founder, Ali Ansari, addressed this issue in a post on X last month, stating that the company does not sell its data to Chinese model makers. Ansari wrote, "Some human data companies work with foreign adversaries. [A]nd the results show today in Kimi K3. We believe it’s shameful to claim American AI dominance [while] selling millions worth of data to countries that we are in adversarial competition with."

Micro1 began as an AI recruiting startup but pivoted into data labeling after clients used its platform to vet and recruit engineers for annotation tasks. The company also builds datasets for robotics pre-training by having generalists record everyday object interactions in their homes. Micro1 raised its Series A at a $500 million valuation in September of the prior year, and TechCrunch reported that the startup may have recently raised another round at a significantly higher valuation.