Micro1
Physical AI
AI data lab that interviews domain experts and sells labeled training data to frontier model builders.
micro1 operates a data lab that finds, interviews and pays domain experts (engineers, doctors, lawyers and others) to create training data and evaluate agents for frontier AI models, including a robotics-video annotation line. The company started as an AI recruiting/interview product and pivoted into training-data supply. Its public site describes an AI interviewer used to vet experts at scale. Founder Ali Ansari started the company in 2022. Press (Forbes, September 2026) reported a raise of more than $100 million at a $4 billion valuation and very large ARR claims; those figures are press-reported, not company-filed, and are not treated here as verified financials. No named lab customers are disclosed on the public site.
Research mapping
Physical AI / AI Factory & Cloud Training Infrastructure (L1) / B
Evidence boundary
Original analysis and independent corroboration remain under development. Commercial scale and durable technical advantage remain under review.
What makes it defensible
Expert-sourced training data via an AI interviewer — the human-data bottleneck model labs hit after scraping and cheap clickwork run out.
Position in the stack: Data supplier to foundation-model labs, now including robotics annotation. Switching costs are relationship and workflow embedding with labs, not a unique sensor; quality of expert panels is the claimed edge.
Who buys it
Founding team
Ali Ansari — Founder and CEO — Solo founder; started micro1 in 2022 (LinkedIn: April 2022) after moving from Tehran to Los Angeles as a child and attending UC Berkeley; pivoted the company from AI recruiting into a training-data lab
SourceFinancing detail
- Last round
- Series B
Investors: Jason Calacanis; Joshua Browder; 01 Advisors
Mapped funds and investors
Direct relationships supported by published evidence.
Reported holdings as-of
No sourced relationships mapped yet.
Disclosed investment history
No sourced relationships mapped yet.
Showing up to four relationships in each group. ETF weights describe the fund’s portfolio, not ownership of this company. Private investment history does not establish current ownership.
Evidence & media
-
This 25-year-old raised over $100M for his AI data startup at a $4 billion valuation
What this source supports: micro1 raised >$100M at a $4B valuation; sells human and synthetic training data to frontier AI labs; gross run-rate ~$500M as of Aug 2026
What it does not establish: Valuation is investor/press-reported, not company-confirmed; gross run-rate is not recognized revenue and is unaudited
Source -
AI data startup micro1 reaches $500M gross run rate amid AI training boom
What this source supports: micro1's gross annual run rate grew from $100M to $500M in eight months, driven by synthetic data and reusable datasets
What it does not establish: Run-rate is attributed to 'a person familiar'; no audited financials and no named customers
Source -
micro1 raises $100M at $4B valuation for AI training data
What this source supports: Deal record: micro1 raises $100M at a $4B valuation to scale AI training data
What it does not establish: Aggregator record; no primary confirmation of terms
Source -
micro1 Announces $1 Billion Enterprise AI Data Commitment
What this source supports: micro1 commits $1B over 12 months to acquire/license de-identified enterprise operational data via its Company Data Partnerships program, building reinforcement-learning environments on real business workflows. Financed by capital from Citi and Hercules Capital.
What it does not establish: First-party release. No named data suppliers, no signed data-partnership roster, no revenue or margin detail on the program.
Source