Mecka AI Raises $60M to Train Robots With Human Data

Mecka AI, a New York City-based robotics startup, announced it has secured $60 million in cumulative funding, including a $25 million Series A round from November 2025 and a $35 million follow-on investment. The round was led by Framework Ventures, with participation from Menlo Ventures, Kindred Ventures, SV Angel, and robotics researcher Ted Xiao, formerly of Google DeepMind. The company differentiates itself by collecting real-world human movement data instead of relying on teleoperation or digital simulations. Mecka’s approach involves deploying a global network of contributors equipped with wearable body sensors and smartphones to capture first-person multimodal data. This data is compiled into Egoverse, a proprietary dataset mapping human interactions with physical environments, including homes, industrial kitchens, and laboratories. Mecka’s proprietary technology processes raw data into precise 3D spatial reconstructions and motion tracking pathways, serving as a training blueprint for robotics hardware. By analyzing Egoverse, developers can train neural networks to replicate human-like movements, such as balancing, reaching, and navigating complex spaces, accelerating commercial robotics deployment. Founded in 2025 under CEO Josh Gao, Mecka claims an annual recurring revenue run rate of $100 million from undisclosed enterprise robotics customers. The new funding will expand its 40-person team, particularly in data engineering and machine learning, while scaling global data collection operations, including hubs in the Philippines. The company’s goal is to establish itself as a critical data infrastructure provider for embodied AI, addressing limitations in traditional robotics training methods. By leveraging real-world human data, Mecka aims to bridge the gap between lab prototypes and commercially viable autonomous systems.