Innodata opens motion-capture lab to train humanoid robots with sub-millimeter data

Humanoid robots can already move with considerable agility, but the AI that drives them still lacks the data it needs to perform useful work. Innodata Inc. announced today that it has opened a laboratory dedicated to generating training data for the next generation of human-like machines.

The New Jersey facility will also independently validate the performance data that robots generate internally, giving developers an external check on their own measurements.

"Physical AI is growing faster than any other segment in AI, but every robotics team hits the same wall: there isn't enough real-world interaction data, and what exists is expensive and slow to produce," said Rahul Singhal, Innodata's CEO.

He said the facility removes that wall, and that Innodata now offers physical AI companies an end-to-end capability spanning data collection through model evaluation — compressing development cycles and bringing more capable, safer robots to market sooner.

Founded in 1988, Innodata argues that data and AI are inextricably linked. The Ridgefield Park, N.J.-based data engineering company said it supplies the high-quality data, evaluation frameworks, and human expertise needed to build AI systems that builders and adopters can trust at scale.

Large language models benefited from the vast amount of text on the internet, but robots have no equivalent corpus describing the physical world, noted Franklin Tanner, Innodata's vice president of robotics and physical AI.

He told The Robot Report that physical AI must "earn its tokens one interaction at a time," and that those interactions have to be deliberate.