समाचार Samāchār · Startups · 19 November 2026

Humyn Labs raises $47 million to film real work for robots. India sells the data layer

The Bengaluru startup, founded in February, raised $26 million in equity led by Accel and $21 million in venture debt. It records people doing everyday tasks in more than 20 countries and sells the footage to AI labs and robotics companies.

A Bengaluru company that did not exist in January has raised $47 million. Humyn Labs records people doing real work and real chores. It labels the footage and sells it to the labs that train robots. The round was announced on 18 November.

The money comes in two parts. Accel led $26 million of equity. Prosus Ventures, BeeGlobal, Bajaj Finance’s Finserv Intelligence, Bajaj Alts AI Opportunities Fund, Dallas Venture Capital and Jump Trading joined, according to TechCrunch and Business Standard. Lighthouse Canton led $21 million of venture debt with Trifecta Capital. The debt is for working capital. Humyn pays its collectors daily or weekly. Its customers pay in 30 to 60 days.

The founders are Manish Agarwal, former CEO of Nazara Technologies, and Ishank Gupta, a former AB InBev executive. They earlier built KGeN, whose contributor network helps Humyn reach homes. Humyn built its own two-camera rig. It collects in more than 20 countries across Latin America and Asia, with Brazil, India and South Korea as hubs. It has 51 full-time staff, mostly in Bengaluru.

The numbers are early but real. TechCrunch reports $10 million of revenue between April and October, $20 million more in signed contracts and 12 paying customers. About 90% of revenue comes from the United States. Business Standard reports an annualised run rate above $50 million. The company says it will build deployment test centres in Brazil, South Korea and India and grow teams in San Francisco, London and Seoul.

This is services export in a new form. India once sold code and back-office hours. Now it can sell the human signal that physical AI needs. The next step is harder. Data is the input. Testing robots for deployment is where the margin will be. Agarwal says as much himself.

What a founder takes from it: Robots will learn from human hands, and India has the people, the field networks and the operators to collect that data at scale. The value is in moving up from footage to deployment testing before others do. Consent and fair pay for the people on camera must be built in from day one.