Sequoia Capital led a $60 million round into Mecka AI, a company that records people performing ordinary physical tasks and turns that footage into structured motion data for training robots. The approach addresses one of the thorniest bottlenecks in humanoid robotics: machines can increasingly be built with sophisticated limbs and sensors, but teaching them to fold laundry, load a dishwasher, or navigate a cluttered kitchen requires enormous amounts of real-world movement data that simply doesn't exist in usable form yet.
Mecka AI's model is straightforward but labor-intensive. The company pays individuals to perform common household and workplace tasks while being recorded, then processes that footage into datasets that robotics companies can use to train their machines' physical behavior. It's a twist on the data-labeling industry that powered the last decade of AI progress, except instead of annotating images or text, Mecka is capturing the physics of human motion itself.
The timing matters. Humanoid robot developers, from well-funded startups to large manufacturers, have poured resources into hardware, but many have openly acknowledged that software and training data lag behind. Unlike language models, which could be trained on the vast text and image archives of the internet, robots learning to move through physical space have no equivalent trove to draw from. That scarcity has turned motion-capture data into a commodity worth competing over, and investors are clearly betting that whoever controls the best datasets could hold outsized leverage as the humanoid robot market matures.
Sequoia's involvement adds weight to that bet. The firm has a long track record of identifying infrastructure plays that sit beneath flashier, more visible technology trends, and data pipelines for robotics fit that pattern: less glamorous than the robots themselves, but potentially indispensable to every company trying to build them. A $60 million check is substantial for a company operating in a data-services niche, signaling confidence that demand from robotics makers will scale quickly.
What happens next will likely hinge on how many robotics firms sign on as customers and how defensible Mecka's data actually proves to be. Competitors could attempt similar human-recording schemes, and some robot makers may try to build in-house motion datasets rather than outsource the work. But if humanoid robots are going to move beyond scripted demos and into homes, warehouses, and factories, someone has to supply the raw material for teaching them how humans actually move through the world. Mecka AI is wagering it can be that supplier, and Sequoia just backed the wager.