AI Training-Data Startup Micro1 Reportedly Hits $4B Valuation

The AI training-data market that Scale AI once defined was upended last June. Meta invested $14.3 billion in Scale AI and hired away founder Alexander Wang, and OpenAI and Google — both major Scale AI customers — began pulling back, saying they needed neutral data partners. Among the challengers racing to fill that gap, none has climbed faster than Micro1.

The San Francisco-based startup connects AI labs with domain experts — doctors, lawyers, software engineers — to produce the high-quality human data needed to train frontier models. Its in-house AI interviewer, Zara, screens and vets every candidate before they join the network. Micro1’s customer list includes frontier AI labs along with Microsoft, Amazon, and robotics firm 1X. The company has also been expanding into synthetic data that doesn’t require human involvement, selling the same “off-the-shelf” datasets to multiple customers at gross margins reportedly as high as 80-90%. That practice has drawn criticism for potentially helping Chinese AI developers catch up; founder Ali Ansari has said publicly that Micro1 doesn’t sell its data to Chinese model makers.

Micro1 was founded by Ali Ansari, who started his first venture at age 12. He built software agency moontek while at UC Berkeley, which became the foundation for Micro1. The company began as an engineering-recruitment business before pivoting entirely to supplying frontier AI labs with high-quality human data.

According to multiple media reports, Micro1 has recently raised more than $100 million at a $4 billion valuation. That’s an eightfold jump from the $500 million valuation it reached with a $35 million Series A in September 2025. Two frontier AI labs that are also Micro1 customers, along with two co-founders of xAI, reportedly participated as individual investors in the round. Neither Micro1 nor the investors have made an official announcement.

The growth numbers back up the valuation jump. Micro1’s annualized revenue grew from $7 million in early 2025 to more than $500 million today. According to TechCrunch, gross annualized run rate alone jumped from $100 million to $500 million between December and August — an eight-month span — though net revenue, after the roughly 60-70% paid out to contracted experts, is estimated at $150-200 million.

Competition keeps intensifying. Mercor, which raised $350 million at a $10 billion valuation last October, has since crossed $2 billion in gross annualized revenue this summer to cement its decacorn status. Bootstrapped Surge AI posted $1.4 billion in revenue in 2025 and, per Bloomberg, is reportedly pursuing its first-ever outside funding — $1 billion at a $25 billion valuation — though that hasn’t been confirmed. Malaysia’s sovereign wealth fund backed Turing, now valued at $2.2 billion. And Handshake, which refounded itself from a college recruiting platform into an AI data supplier, has topped $1.1 billion in gross annualized revenue.