
Maxinsights raises $30 million for physical AI training data
The AMW Read
The seed round adds a supplier and reported collection scale to the physical AI data market, but its broader effect remains unproven.
Maxinsights raises $30 million for physical AI training data
Maxinsights raised $30 million in seed funding to expand its platform for collecting training data for robots and AI models. Peak XV led the round, with participation from Tola Capital, South Park Commons, INCE Capital, and Alumni Ventures. The company pays contractors to record everyday activities and says it has captured more than 2 million hours, including 1.5 million hours it considers model-ready. It reports serving more than 30 customers, including AI labs and foundation model developers.
The funding places Maxinsights in the data infrastructure layer for physical AI, where suppliers must turn varied real-world recordings into material model developers can use. Its network includes more than 5,000 contractors across six countries, recording tasks in homes and workplaces. Maxinsights says its MaxLoop system assesses incoming footage and provides feedback to operators. The company claims the system reduced review costs by about 80% and can process 450,000 hours of footage per month. These figures are company claims, not independent measures of model performance.
For builders and investors, the test is whether collection quality and additional sensor data improve training outcomes enough to justify their cost. Maxinsights is developing MaxGlove with Genesis AI to capture force and touch information that video alone cannot provide. The company targets 1 million hours of force data in 2027. It also says it supplied all 200,000 hours of egocentric data used to train Genesis AI's GENE-26.5 model and is the largest data provider behind Dyna Robotics' Dyna-2 World-Action Model. Those supply relationships do not establish how much the data improved either model.