Your Daily Chores Could Train Tomorrow’s Robots Startup Maxinsights Raises $30 Million”

Webdesk- Shiza Iqbal

Maxinsights Raises $30 Million to Train Robots Using Videos of Everyday Household Chores

In a move highlighting the growing demand for real-world artificial intelligence training data, robotics startup Maxinsights has raised $30 million in seed funding to expand its platform, which collects videos of people performing everyday household tasks to help robots learn human activities

The investment round was led by Peak XV, with participation from Tola Capital, South Park Commons, INCE Capital and Alumni Ventures. Based in Santa Clara, California, the company plans to use the funding to scale its data collection operations and support developers building next-generation robotic systems.

Unlike large language models that learn from vast collections of online text, robots require a different kind of training material: visual information showing how humans interact with physical objects in real-world environments. From preparing meals to cleaning rooms, these routine activities can be difficult for robots to replicate without sufficient examples of human movement and decision-making.

Yuguang Yuan, founder and chief executive officer of Maxinsights, said the company was established to address this shortage by building an end-to-end system for collecting and processing real-world human activity data.

“We built the whole pipeline ourselves, from the hand-tracking models to the platform that scores every operator’s footage as it arrives,” Yuguang Yuan said, explaining the technical challenges involved in making real-world footage useful for robotics developers.

Maxinsights employs more than 5,000 contractors across six countries. These workers wear specialized cameras mounted on their heads, wrists and chests while carrying out ordinary activities, including cooking, washing cars and cleaning homes.

The approach allows the company to capture human movements at natural speeds in everyday settings rather than relying exclusively on controlled laboratory demonstrations. According to the company, this helps robotics developers study unpredictable situations that machines may encounter when operating in real homes.

To date, Maxinsights has collected more than two million hours of footage. Its automated quality-control platform, MaxLoop, processes millions of video clips to assess incoming data and maintain consistency while reducing the cost of manual review.

The company has also partnered with developers to create MaxGlove, a specialized sensor glove designed to measure grip force and object weight alongside visual recordings. Such information could help robots better understand how much force is needed to handle different objects safely.

As investment in physical artificial intelligence continues to grow, access to reliable training data is emerging as a critical challenge for the robotics industry. Maxinsights aims to address that gap by turning ordinary human activities into valuable learning material for machines.

However, the broader expansion of camera-based data collection also highlights the importance of privacy, informed consent and responsible handling of recordings captured inside private homes.

The success of platforms such as Maxinsights may ultimately depend on whether they can provide high-quality training data at scale while maintaining public trust. If successful, the company’s approach could help bring robots closer to performing useful everyday tasks in real-world environments.