Mecka AI nears $500M valuation in Sequoia-led deal amid rush for robot training data

Mecka AI, a startup that gathers and analyzes human motion data to train humanoid robots, is on the verge of a $500 million valuation after securing a new round of investment led by Sequoia Capital, according to insiders. This comes just three months after Mecka raised $60 million in funding led by Framework Ventures, with participation from other notable venture capital firms.

The exact size of the new round remains undisclosed, as the terms of the deal are still pending finalization. The company's founders, Josh Gao, Mogen Cheng, Jason Chong, and Duy Nguyen, joined forces in 2024 with no prior robotics experience. They identified a lack of physical-world data as a major obstacle to developing general-purpose robots, including humanoid models.

To address this, Mecka launched a unique approach of incentivizing individuals to record themselves performing daily tasks using body sensors and smartphones, similar to how data companies Scale AI, Mercor, and Surge have supported the growth of large language models (LLMs). As of early June, Mecka projected reaching a $100 million annual run rate by the end of 2026.

While the company has not disclosed its customers, numerous robotics firms and AI labs rely on the data Mecka collects through this "egocentric" method, along with other physical data collection techniques such as teleoperation, to enhance their models. Competing startups like XDOF and data platforms like Scale AI and Micro1 are also expanding their offerings to cater to the growing demand for real-world data in robot training.