Maxinsights, a Santa Clara, California-based company developing training data infrastructure for physical artificial intelligence, has raised $30 million in seed funding led by Peak XV. The financing included participation from Tola Capital, South Park Commons, INCE Capital, and Alumni Ventures.
Alongside the funding announcement, Maxinsights introduced MaxLoop, an AI-powered data quality and annotation platform, and MaxGlove, a sensor-equipped glove designed to capture tactile information and physical forces during everyday activities.
The company develops datasets used to train robotics foundation models by recording how people perform real-world tasks, including cooking, cleaning, organizing objects, and completing workplace activities.
Unlike approaches relying primarily on simulated or staged environments, Maxinsights collects first-person video footage from participants performing activities in homes, restaurants, hotels, and other everyday settings.
Its collection network includes more than 5,000 contractors across six countries who are compensated for recording activities using the company’s proprietary camera hardware.
Maxinsights has collected more than 2 million hours of video, including approximately 1.5 million hours of model-ready data available for customers.
The company serves more than 30 customers, including frontier AI laboratories and foundation model developers, with eight customers operating under recurring contracts.
Maxinsights’ data infrastructure follows a multistage process involving collection, quality review, annotation, curation, and delivery.
Its proprietary systems analyze footage to identify environmental conditions, tasks, subtasks, instructions, hand movements, body positions, and interactions with objects.
These capabilities help robotics developers train models to understand how people perform physical activities across varying environments and circumstances.
The newly introduced MaxLoop platform automates significant portions of the data quality control process.
Using proprietary hand-tracking and video analysis models, MaxLoop evaluates footage as it arrives, identifying issues involving visibility, movement, task completion, synchronization, and other quality requirements.
The company said its technology has reduced manual review requirements by approximately 90%, lowered review costs by roughly 80%, and increased processing capacity to approximately 450,000 hours of footage per month.
Maxinsights also introduced MaxGlove, a wearable sensor system being developed in collaboration with Genesis AI.
The glove is designed to record information that conventional cameras cannot directly capture, including grip strength, object weight, pressure, resistance, and changes in physical contact.
These measurements are important for developing robots capable of handling objects safely and adjusting their movements based on tactile feedback.
Maxinsights aims to collect 1 million hours of force data in 2027 by deploying MaxGlove across its existing collection network.
The company has already contributed datasets to several advanced robotics development programs.
Genesis AI used 200,000 hours of first-person video supplied exclusively by Maxinsights to train its GENE-26.5 robotics foundation model.
Maxinsights is also the largest data provider supporting Dyna Robotics’ Dyna-2 World-Action Model.
The company was founded by CEO Yuguang Yuan and CTO Kun Wang, who bring experience in robotics, autonomous vehicles, and industrial technology.
Yuan previously co-founded Mega Gear Robotics, an industrial robotics company that was acquired. He also held leadership responsibilities at Eaton involving businesses generating approximately $150 million in annual revenue.
Wang previously worked at Google X and Waymo before co-founding QCraft, an autonomous driving technology company. Maxinsights’ broader engineering team includes professionals with experience at Google, Microsoft, Meta, Waymo, and Hulu.
The funding comes as investment in physical AI continues expanding. Maxinsights cited approximately $47.4 billion invested in the sector during the first half of 2026, highlighting demand for the training infrastructure needed to develop increasingly capable robotics systems.
With its new financing, Maxinsights plans to expand its collection infrastructure, improve automated data processing, and develop additional sensing capabilities that help robots learn from natural human behavior.
KEY QUOTES:
“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. Robots get deployed into real kitchens, not staged ones, and something unexpected happens every few minutes. Edge cases are the job, and the only way to get them is to record real work in real homes. Collecting one great hour of data is easy, and holding quality across millions of hours and thousands of people is what we built the company to do.”
Yuguang (YG) Yuan, Founder And CEO Of Maxinsights
“Almost every company in this category is selling hours. Maxinsights is the only one that treated quality control as an engineering problem rather than a staffing cost, which is what lets them operate in real life. That decision is hard to copy after the fact, and it’s why the labs that tried other vendors keep coming back to them.”
Rohit Agarwal, Managing Director Of Peak XV
“Robots need to learn from the complexity and variation of everyday life, and carefully controlled datasets alone cannot capture that. With Genesis 26.5, we focused on footage of people performing everyday tasks naturally, with much of that data provided by Maxinsights, to help our models learn from a broader range of real-world situations.”
Johnson Wang, Co-Founder Of Genesis AI
