Mecka AI

Mecka AI collects human motion data and sells it to robotics companies and AI labs as training data for humanoid and other robots. It pays people to record themselves performing ordinary physical tasks, such as making coffee or repairing a car, using body sensors and smartphone cameras. The approach is known as egocentric data collection, and it sits alongside teleoperation as one of the two main ways the field gathers real-world physical data. The underlying bet is that robot foundation models are constrained less by architecture than by the scarcity of real physical demonstration data.

The company was founded in 2024 by Josh Gao, Mogen Cheng, Jason Chong and Duy Nguyen. In mid-2026 it raised a 60 million dollar round led by Framework Ventures, with Menlo Ventures, SV Angel and Kindred Ventures participating. In September 2026 TechCrunch reported that Sequoia Capital was leading a further round that would value Mecka at about 500 million dollars, while noting that the terms were not final and that the size of the round was not known. As of early June 2026 the company projected ending that year at an annual run rate of 100 million dollars.

Mecka sits in a fast-forming category of embodied-data suppliers that includes Scale AI, Mercor, Surge, XDOF and Micro1, applying to robotics the data-supply trade those companies built for language and vision models. It sells a data service rather than a product, with no self-serve access and no published pricing.