AI training data market grows rapidly as Micro1 sees $5 billion in revenue

The demand for high-quality AI training data has led to significant growth in data labeling startups, with Micro1 experiencing a sharp increase in revenue. Founded four years ago, the company's annualized revenue has risen from $100 million to $500 million in just eight months, according to insiders. Micro1 typically retains 60% to 70% of contract revenue, resulting in an estimated annual net income of $150 million to $200 million. Currently, Micro1's size is still smaller than Mercor and Handshake, which reported annualized revenues of $2 billion and $1 billion, respectively. However, the rapid growth of Micro1 highlights the expanding demand for AI training data, which some researchers believe could eventually match the scale of spending on computing power. Micro1 is expanding its synthetic data production by automatically generating video descriptions, with some data being resold to multiple customers. This practice has achieved a gross margin of 80% to 90% for 'off-the-shelf' data. Ali Ansari, founder of Micro1, stated that the company will not sell data to Chinese AI developers, citing concerns over potential enhancement of their model capabilities. Originally launched as an AI recruitment company, Micro1 shifted focus to training data after noticing customers used its platform to screen data labelers. The company now includes experts who evaluate model outputs and conduct reinforcement learning training, while also building a robot pre-training dataset. This dataset involves hundreds of ordinary users recording their interactions with everyday objects. In September 2025, Micro1 completed its Series A funding round, achieving a valuation of $500 million. Reports suggest the company may soon complete another round of funding, potentially leading to a significant increase in its valuation.