Micro1, a training-data supplier to frontier AI labs, has raised $100 million at a $4 billion valuation, according to a Forbes report. The company's gross annual run rate has reached $500 million.
The jump marks an extraordinary acceleration for the startup. At the start of 2025, Micro1 was a recruiting firm with $7 million in annualized revenue. The company has since pivoted entirely into AI data, growing its gross run rate 71 times over in less than a year. The valuation climbed from $500 million in September 2025 to $4 billion in the current round, an eight-fold increase.
Micro1 was founded by Ali Ansari, 25. The company supplies training datasets and human-annotated content to AI developers. Its customer base includes Microsoft, Amazon, and 1X, along with unspecified frontier AI labs. The startup has reported gross margins of 60 to 70 percent on its $500 million run rate.
Two cofounders from xAI and two representatives from frontier labs participated in the new funding round, according to the report. Large AI companies are competing intensely for proprietary training data as model scaling has become capital-constrained by data availability rather than compute.
Training-data startups have become central to AI development timelines. Major labs including OpenAI, Anthropic and Google have faced bottlenecks in sourcing human feedback and labeled datasets at the scale required for frontier model training. The scarcity has pushed prices for high-quality annotation work higher and drawn venture capital into the space.
Micro1's growth rate outpaces most early-stage enterprise software companies. A 71-fold increase in ARR within a single year is rare outside of viral consumer products or constrained-supply markets. The company's $500 million gross run rate at a $4 billion valuation implies an implied multiple of 8 times gross revenue, typical for high-margin SaaS but aggressive for a company that pivoted its entire business model nine months ago.
The speed of Micro1's scaling and the participation of xAI executives in its funding show that frontier labs view training-data supply as a competitive bottleneck. If the company sustains its $500 million run rate and margins remain at 60 to 70 percent, the path to profitability at scale is visible; the risk is whether demand for annotation and human feedback stabilizes or compresses as AI companies develop synthetic data generation and in-house labeling infrastructure.
