Robot training data startup Mecka AI has raised $60 million in a Series B round led by Sequoia Capital. NVIDIA and Microsoft’s venture fund M12 also participated in the round. Reportedly, Mecka previously approached a $500 million valuation.
Founded in 2024, Mecka pays people to record themselves doing everyday tasks. Workers wear body sensors and use smartphones to capture actions like making coffee or fixing cars. The startup then processes this human motion data to train humanoid and other robots. In short, Mecka wants to do for robotics what Scale AI and Mercor did for large language models.
Four co-founders lead the company. CEO Josh Gao and Mogen Cheng previously built a restaurant fintech startup in Canada. Jason Chong ran a crypto exchange that Coinbase later acquired. Duy Nguyen handles operations. Notably, none of the founders come from a robotics background.
However, the team identified a critical bottleneck early. General-purpose robots need massive amounts of real-world interaction data to function reliably. Mecka uses an “egocentric” approach, capturing tasks from the performer’s own point of view. This method differs from teleoperation, where a human remotely controls a robot to generate training data.
The market for robot training data is heating up fast. Competitor XDOF reportedly nears a $1.2 billion valuation for its own Series B. Meanwhile, Scale AI and Micro1 now also expand into robotics data. As a result, the space has become one of the fastest-growing segments in AI infrastructure.
Mecka projects an annual revenue run rate of $100 million by the end of 2026, according to comments Gao made to Fortune. The company plans to use the new funding to scale its data infrastructure and expand into additional industry verticals.
Gao has argued that robotics is nearing an inflection point. Better models, more capable hardware, and growing commercial demand all drive this shift. Yet the key challenge remains the same: scaling real-world experience for machines that must operate in unpredictable environments.
