Maxinsights raises $30 million to scale physical AI data platform

Maxinsights is betting that the hardest part of physical AI is not collecting video but keeping that video usable at scale. The company says it built its entire pipeline in-house, from hand-tracking models to a platform that scores every operator's footage as it arrives. That infrastructure claim is central to its pitch: collecting one great hour of data is easy, but holding quality across millions of hours and thousands of people is the operational problem it aims to solve.

The funding comes alongside two new tools. MaxLoop automatically assesses incoming footage and gives feedback to operators; Maxinsights says it has cut review costs by approximately 80 percent and enables processing of 450,000 hours of footage a month. MaxGlove, developed with Genesis AI, captures tactile information such as grip force and object weight that video alone cannot provide. The company says all 200,000 hours of egocentric data used to train Genesis AI's GENE-26.5 model came from Maxinsights, and that it is the largest data provider behind Dyna Robotics' Dyna-2 World-Action Model.

The approach contrasts with staged data collection. By recording real homes and workplaces, Maxinsights aims to expose AI models to the variation and unexpected events that robots encounter in deployment. With more than 5,000 contractors across six countries and over 30 customers, the company is now targeting 1 million hours of force data in 2027, a signal that touch, not just vision, is the next frontier it intends to supply.