Mecka raises $60M Series B to digitize human motion for robot training

In short: Mecka closed a $60 million Series B led by Sequoia Capital on Oct 7, 2026, with new capital from NVIDIA, M12 (Microsoft's venture fund), Qualcomm Ventures and Samsung. The company builds capture hardware and reconstruction models so robots can learn from real human motion — and says it crossed $100M run-rate revenue in June 2026.

What Mecka sells

• Instruments: multi-sensor capture hardware designed in-house for the signals downstream models need.
• Capture fleets: people recording everyday tasks (coffee, car repair) with body sensors and phones.
• Research stack: motion tracking, 3D reconstruction and sub-centimeter hand-pose accuracy on in-the-wild data.
• Deployment: Mecka also positions itself as a robotics integrator that captures on-site data, post-trains, and improves with every hour of operation.

Investors also include Kindred, Framework Ventures, Neo, and angels such as DoorDash CEO Tony Xu and former Optimus head Milan Kovac. TechCrunch had earlier reported talks near a $500M valuation.

Why it matters

Foundation-model labs still starve for contact-rich human demonstration. Mecka is betting that imitation learning at Scale-AI-like volume — not more scraped video alone — is the bottleneck for embodied AI and humanoid robots.