NYC-based Mecka, which collects human motion data to train humanoid robots, raised a $60M Series B led by Sequoia; Nvidia, M12, Samsung, and others invested too
Mecka AI, a startup that collects and analyzes human motion data to train humanoid robots and other kinds of robots …
Context & Ripple Effects
June coverage described Mecka’s earlier financing, including a $25 million Series A, around its use of body sensors and iPhones to source training data. The new round puts a larger investor group behind the data-collection layer rather than a humanoid maker alone.
Mecka is part of a widening race to build proprietary real-world robot datasets: Munich’s Microagi raised $55 million for factory and household data collection, while Figure AI had already raised $675 million for a general-purpose humanoid robot.
First-order effects
- Mecka gains $60 million and backing from Sequoia, Nvidia, M12 and others to expand its human-motion-data operation for robot training.
- Nvidia becomes an investor in a robotics-data supplier, extending its exposure beyond the robot makers that consume training and compute resources.
Second-order effects
- Microagi and other robot-data collectors face a better-capitalized Mecka in competing for data-collection capacity and customers building humanoid systems.
- Humanoid developers such as Figure AI have another potential specialist source of training data, separating data acquisition from robot design and manufacturing.
Third-order effects
- If funding continues to favor specialist data collectors alongside humanoid makers, proprietary real-world datasets may become a distinct control point in the robotics supply chain rather than an in-house input.
- The overlap of venture firms, strategic chip investors and robotics startups signals an AI capital stack extending from compute into the physical data needed to train embodied systems.
The trend: Humanoid robotics is attracting investment not only into robots themselves, but into the specialized data pipelines required to train them.