Mecka AI is receiving a new investment of approximately $500 million from Sequoia Capital. This follows a previous $60 million round led by Framework Ventures, which included participation from Menlo Ventures, SV Angel, and Kindred Ventures. The company’s focus is on collecting physical-world data to train robots, specifically humanoid robots. This data is captured through payments to individuals performing everyday tasks using body sensors and smartphones.
The startup was founded in 2024 by four entrepreneurs, including Canadians Josh Gao and Mogen Cheng, and Jason Chong, who previously worked at Coinbase. The team’s initial recognition of a shortage of physical-world data identified a key obstacle to the development of general-purpose robots. Mecka’s approach mirrors that of companies like Scale AI and Mercor, which provide data for training large language models.
As of early June 2026, Mecka AI projected an annual run rate of $100 million. The company’s methodology involves paying individuals to record themselves performing tasks such as making coffee or fixing cars. This ‘egocentric’ data collection approach is used by robotics companies and AI labs alongside other physical data collection methods.
Other companies in the real-world data collection space for robotics include XDOF, which is nearing a $1.2 billion valuation, and Scale AI, Micro1, which are expanding beyond LLMs. Mecka AI’s strategy aims to address the significant data gap hindering the advancement of robotics technology.
