In 2024, Mecka AI, a New York-headquartered startup, was founded by four entrepreneurs who had no prior experience in robotics. By September 2026, the company is on the verge of finalizing a new funding round led by Sequoia Capital, with its valuation estimated at around $500 million. Unlike traditional robotics firms, Mecka AI doesn't manufacture robots or train models directly. Instead, it hires everyday individuals to wear its proprietary sensors and haptic gloves, using iPhones to capture real-world work processes in real-time. The multi-channel synchronized data gathered is then transformed into robot-learnable data and stored in the Egoverse library, which has amassed over 30,000 hours of content. The crux of the challenge lies in translating human motions into robotic actions. Mecka AI's business model encompasses selling tailored datasets, along with providing subsequent robot deployment and maintenance services. In June 2026, the company revealed an annualized recurring revenue run rate of approximately $100 million, with 1X Technologies as its sole confirmed paying customer. In contrast, the U.S.-based Figure adopts a paid crowdsourcing approach for data collection, while China's JD Cloud plans to gather data through home services, sparking compliance concerns. Other players like Zhiyuan and Unitree primarily collect data in their self-built facilities and make it openly available. Domestic data collection efforts are predominantly project-driven, making it challenging to create reusable assets. While Mecka AI's model allows for asset accumulation through client projects, questions linger about the technical robustness of transferring human movements to robots, and the industry grapples with inconsistent data standards.