AI training data demand is reshaping the market for data-labeling startups, and Micro1 has emerged as one of the fastest climbers in the field. The four-year-old company has increased its gross annual run rate from $100 million to $500 million in just eight months, according to a person familiar with the business.
Like other firms that rely on contract specialists such as doctors, lawyers, and scientists, Micro1 keeps about 60% to 70% of that total, placing its net annual run rate at roughly $150 million to $200 million. While rivals such as Mercor and Handshake have posted even larger figures, Micro1's momentum suggests the market can support several players at once.
The company is also leaning more heavily into synthetic data, including automated descriptions of video content. Some of these datasets can be sold to multiple customers, with gross margins reportedly reaching 80% to 90% for off-the-shelf products. That shift reflects a broader industry trend toward scalable, reusable training assets.
Micro1 began as an AI recruiting startup before moving into data labeling after seeing how clients used its platform to find annotation talent. Founder Ali Ansari has also said the company is building robotics pre-training datasets through everyday object-interaction recordings collected from generalists at home.
After raising a Series A at a $500 million valuation last year, the startup may have secured another round at a higher valuation. If AI spending on data continues to rise, companies like Micro1 could help define the next phase of model development and data infrastructure.
