Mecka AI raises $60 million to teach humanoid robots how humans move

# Mecka AI Raises $60 Million to Build Human-Movement Data Business for Robotics

Mecka AI, a startup specializing in collecting and analyzing human-movement data for robotics training, has raised $60 million in a Series B funding round as demand increases for real-world information to develop humanoid robots and other robotic systems.

The funding round was led by Sequoia, with participation from Nvidia, Microsoft’s venture fund M12 and other investors. The investment reflects growing interest in the data infrastructure required to help robots learn and perform tasks in real-world environments.

TechCrunch previously reported that Mecka AI was nearing a new funding round at a valuation of approximately $500 million. The company, founded in 2024, is developing a business focused on capturing how people move, interact with objects and carry out everyday activities.

Unlike companies that collect text, images or other digital information to train artificial intelligence models, Mecka AI concentrates on physical actions. It pays participants to record themselves performing activities such as preparing coffee, repairing vehicles and completing routine household or workplace tasks.

Participants can use smartphones and wearable body sensors to document their movements and interactions with their surroundings. The resulting datasets can help robotics developers understand how humans carry out tasks, providing examples that machines may use to learn similar behaviors.

Such data could become increasingly valuable as robotics companies work toward deploying machines in homes, factories, warehouses and other environments built around human activity. While advances in hardware and AI models remain important, robots also require extensive examples of physical tasks to operate effectively outside controlled settings.

Mecka AI is entering a growing market in which specialized companies are building data collection businesses for robotics. XDOF, another startup focused on gathering real-world data for robot training, was previously reported by TechCrunch to be discussing a Series B funding round at a valuation of $1.2 billion.

Established AI data companies are also expanding into robotics. Scale AI and Micro1, which have worked on supplying data for artificial intelligence model development, have broadened their activities into areas involving robotic systems. Their expansion underscores the growing overlap between AI data services and the development of machines capable of interacting with the physical world.

For robotics developers, gathering useful real-world data remains a significant challenge. Simulations can help train robotic systems, but data collected from people performing actual tasks can offer examples of movement, object handling and environmental interactions that are difficult to reproduce fully in virtual settings.

As technology companies compete to build humanoid robots capable of handling more complex activities, demand for high-quality physical-world datasets is likely to remain an important part of the industry's development. Mecka AI's latest funding provides additional capital to expand its data collection operations and strengthen its position in the emerging robotics data market.