Mecka AI raised $60M to train robots on human motion data
Fortune reports the New York startup closed a $25 million Series A and a $35 million follow-on led by Framework Ventures, while CEO Josh Gao says signed contracts point to a $100 million annual run rate.
By Ryan Merket · · updated
Why it matters
Physical AI startups are trying to turn robotics from demos into deployed systems. Mecka AI is framing its $60 million raise around the less glamorous but critical layer: data, evaluation and deployment infrastructure.
Fortune's Ben Weiss reports that Mecka AI has raised $60 million across two previously unannounced financings to train robotics systems on human movement data collected through body sensors, iPhones and custom hardware.
The New York company closed a $25 million Series A in November and a $35 million follow-on investment, according to Fortune. Framework Ventures led the rounds, with participation from Menlo Ventures, SV Angel, Kindred Ventures and angel investor Ted Xiao, a former Google DeepMind researcher and founding member at Jeff Bezos's Project Prometheus. CEO Josh Gao declined to disclose the valuation.
Gao and cofounders Mogen Cheng, Jason Chong and Duy Nguyen are making a contrarian bet in physical AI: that robots should be trained on data collected from humans, rather than relying primarily on teleoperation, where people manually operate robots. Gao told Fortune the team developed that thesis after studying robotics research, visiting labs and talking with roboticists, then partnered with robotics labs to test whether human-collected data could work at scale.
The founding team did not come out of the usual AI lab pipeline. Gao and Cheng previously sold a restaurant payments technology startup. Chong sold a startup to Coinbase. Nguyen, Gao told Fortune, made millions flipping sneakers. The quartet started Mecka AI in 2025 after collecting what Gao described as thousands upon thousands of gigabytes of data.
Gao says demand is already material: Mecka AI projects a $100 million annual run rate based on signed contracts, though he declined to name customers. Framework cofounder Vance Spencer told Fortune it is the fastest-growing revenue company his firm has backed.
The financing sharpens Mecka AI's pitch from Gao's X thread: physical AI needs a data and deployment layer, not just better models or robot hardware. Gao said the money will go toward scaling data infrastructure, expanding into new verticals and deploying robots into real-world commercial settings.
The unanswered questions remain the same ones facing the broader robotics boom: which customers are paying, which deployments are live and how quickly human-motion data can translate into reliable robots outside controlled environments. Gao's view is more aggressive. "Useful robots can be deployed today," he told Fortune, "and immediately."