Mecka Raises $60M to Feed Humanoid Robots the One Thing They Lack: Real Human Motion
Sequoia, Nvidia, and Microsoft's M12 are betting that the bottleneck in physical AI isn't hardware — it's training data. Mecka just became the clearest proof of that thesis.
The humanoid robot industry has a data problem — and it's not the kind you can solve by scraping the web.
On October 7, 2026, Mecka closed a $60 million Series B led by Sequoia Capital, with participation from Nvidia and Microsoft's M12. The round positions Mecka as one of the most prominent recent financings focused squarely on data infrastructure for physical AI — not on building robots, but on supplying the raw material that makes them trainable.
The Bottleneck Sequoia Is Paying to Solve
Humanoid robots learn to move by studying movement. The challenge is that human motion — the way a hand grips an irregular object, the micro-adjustments a torso makes when carrying uneven weight — is extraordinarily difficult to synthesize from first principles. Simulated data can get a robot to walk. It rarely gets a robot to work.
Mecka's business is built on collecting real-world human-motion data and making it available as training input for humanoid robot developers. That positions the company at the least glamorous and most necessary layer of the physical AI stack: the data layer. Sequoia leading the round — and Nvidia and M12 co-signing it — signals that the people closest to where this industry is heading believe the data infrastructure layer is where durable value accumulates.
Why the Investor Mix Matters
Sequoia provides the capital and the growth-stage credibility. But the strategic weight here comes from the co-investors. Nvidia's presence is not incidental — the company has a direct commercial interest in seeing humanoid robotics scale, given its hardware sits at the center of physical AI compute. Microsoft's M12, the corporate venture arm, adds an enterprise distribution angle. Together, the cap table reads less like a speculative bet and more like a coalition of stakeholders who need the data problem solved.
For Mecka, that backing does something beyond funding: it signals to robot developers that the company's data pipeline has been stress-tested by parties with real technical and commercial exposure to the space.
What the $60M Actually Buys
Mecka plans to deploy the $60 million to expand its robotics training-data business. The specifics of that expansion — new motion-capture infrastructure, geographic coverage, dataset categories — weren't disclosed. But the strategic logic is straightforward: training-data businesses are volume games. More diverse, higher-fidelity human-motion data means more capable models, which means stickier customers among robot developers racing to ship capable hardware.
The round also arrives at a moment when the physical AI category is drawing serious capital, and the competition for quality training data is intensifying. Locking in a strong data position now — before the humanoid hardware wave fully crests — is the play.
The Bigger Shift
The meaningful pattern here isn't a single funding round. It's that serious institutional capital is now flowing into the infrastructure layer of physical AI, not just into the robots themselves. Software AI's last decade taught the market that the companies controlling training pipelines — the data, the labels, the feedback loops — often outlast the model developers they serve.
The humanoid robot industry appears to be absorbing that lesson early. Mecka's $60 million Series B is evidence that the race for physical AI isn't just about which robot walks best. It's about who owns the data that teaches it how.