Mecka AI raises $60 million Series B led by Sequoia

Mecka AI is trying to build the robotics equivalent of the data-labeling layer that companies like Scale AI, Mercor, and Surge provide for LLMs (AI that understands and generates text). Where those firms supply human-generated data for text systems, Mecka pays people to record themselves making coffee or fixing cars while wearing body sensors and holding smartphones, producing the motion data robots learn from.

The company was founded in 2024, so this Series B came together quickly. The article also notes that other startups collecting real-world data for robot training, such as XDOF, are raising at high valuations, while human-data platforms that began with LLMs, including Scale AI and Micro1, are expanding into robotics.

The round's outcome appears to hinge on how much robot makers rely on paid human motion data rather than other training approaches. For Mecka, the test may be whether it can turn its contributor network into durable, hard-to-copy data at scale, since the article does not say how the company performs against those rivals.