Maxinsights raises $30M seed to feed robots real-world training data
What's the deal? MaxinsightsDealroom has a profile for this one. Try Dealroom → has raised $30 million in seed funding to build what it calls the data layer for physical AI. PeakXV led the round, with Tola CapitalDealroom has a profile for this one. Try Dealroom →, South Park Commons, INCE CapitalDealroom has a profile for this one. Try Dealroom →, and Alumni Ventures participating.
How it works: The company pays more than 5,000 contractors across six countries to wear its camera hardware while they cook, clean, and do their jobs. It then turns that first-person footage into training data for robots, and says it has captured more than two million hours, with 1.5 million model-ready for delivery.
Why now? Investment in physical AI is surging — one estimate put sector funding at $47.4 billion in the first half of 2026, nearly four times the second half of 2025 — but usable real-world data remains a bottleneck. Robots learn tasks such as folding a shirt only by watching people repeat them thousands of times, and little work has been captured at the scale models require.
By the numbers: The round ranks in the 99th percentile of seed rounds over the trailing 48 months, based on a sample of 47,814 deals. That puts Maxinsights among the largest seed raises in its region.
What's new: Alongside the funding, Maxinsights launched two products. MaxLoop scores and corrects data as it arrives, cutting its human review team by 90%, while MaxGlove, a sensor glove, records grip strength and object weight that a camera cannot see.
What's the endgame? Maxinsights serves more than 30 customers, including frontier AI labs, with eight on recurring contracts. It is developing MaxGlove with Genesis AI to add touch and force data.
Reality check: The scale and customer figures come from company materials. The round's significance will hinge on whether the data measurably improves robot performance outside controlled settings.
What they're saying: "We built the whole pipeline ourselves, from the hand-tracking models to the platform that scores every operator's footage as it arrives, because nobody was going to sell us the pieces that make messy real-world data usable," said founder and chief executive officer Yuguang Yuan.
The signal: Capital is chasing the physical AI data gap. Rivals including Humyn LabsDealroom has a profile for this one. Try Dealroom →, which said it would deploy $20 million for human-data collection, alongside synthetic-data providers, are racing to supply what robot models need most — and whoever cracks usable real-world data at scale stands to own a critical piece of the stack.
Read more: business.thepilotnews.com
Image credit: Generated with Gemini
