Data-labeling startup Micro1 hits $500M gross run rate as AI training demand surges

Micro1, a data-labeling startup supplying training material to major AI labs and corporations, has dramatically expanded its business, growing its gross annual run rate from $100 million to $500 million over the past eight months. The surge reflects explosive demand for high-quality datasets needed to train AI models.

The opportunity landscape#

The hunger for specialty training data—especially from domain experts like physicians, lawyers, and scientists—is fueling growth across the entire sector. Micro1's trajectory, while impressive, still trails larger competitors: Mercor reached $2 billion in gross annualized revenue this summer, and Handshake hit $1 billion earlier in 2026. Yet the market is expanding fast enough to sustain multiple players, suggesting the boom is far from saturated.

After retaining roughly 60–70% of gross revenue to cover costs, Micro1's net annual run rate stands between $150 million and $200 million. Looking ahead, some researchers expect AI spending on data will eventually match or exceed spending on compute infrastructure—a shift that could accelerate growth further.

Margin expansion and synthetic data#

The company is improving profitability through two key strategies. It is increasingly generating synthetic data without human annotation, such as automatically captioning video content. Additionally, datasets that can be resold to multiple customers carry much higher margins—up to 80–90%—compared to custom work.

This off-the-shelf data model has become controversial, however. Critics have flagged that selling identical datasets to Chinese AI developers undermines U.S. competitive advantage. Micro1's founder Ali Ansari has publicly distanced the company from the practice, stating on X that the startup does not sell its data to Chinese model makers. "Some human data companies work with foreign adversaries," Ansari posted. "We believe it's shameful to claim American AI dominance desires while selling millions worth of data to countries that we are in adversarial competition with."

From recruiting to data#

Micro1 was originally an AI recruiting platform before Ansari pivoted into data labeling. He noticed that his platform's data-labeling capabilities were being used by clients to hire and vet annotators, signaling where the real demand lay. The company now focuses on work ranging from reinforcement learning evaluations—where experts assess model outputs—to robotics pre-training, achieved by having hundreds of contract workers record everyday object interactions at home.