Mecka AI raised a $60 million Series A to scale a platform that trains robots using human data, according to Fortune. Rather than relying solely on teleoperation or simulation, Mecka captures real human motion through wearable body sensors and consumer iPhones, turning everyday physical activity into demonstration data that robots can learn from.
The thesis targets what many in the field now see as the true bottleneck in embodied AI: not the robot bodies or the foundation models, but the volume and quality of real-world data needed to teach machines how to move and manipulate objects reliably. Crowdsourced human demonstration, captured cheaply on devices people already own, is a potential shortcut around expensive lab data collection.
“Crowdsourced human demonstration, captured cheaply on devices people already own, is a potential shortcut around expensive lab data collection.”
The round fits the broader robotics funding record, but from a different angle -- the data layer rather than the hardware or model layer. As capital floods into humanoids and general-purpose robots, infrastructure that makes those systems actually learn faster could prove to be a uniquely defensible position.