November 19, 2026Funding-Series ARobotics

Humyn Labs Raises $47M Series A to Film Human Chores for Robot Labs

Humyn Labs, a Bengaluru startup founded only in February, raised a $47M Series A, but split it in two before quoting it: $26M is equity led by Accel, with Prosus Ventures, BeeGlobal, Bajaj Finserv Intelligence, Bajaj Alts AI Opportunities Fund, Dallas Venture Capital, Lighthouse Canton and Jump Trading; $21M is venture debt led by Lighthouse Canton with Trifecta Capital. No valuation was disclosed. TechCrunch: https://techcrunch.com/2026/11/18/accel-backs-indias-humyn-labs-to-supply-training-data-needed-by-u-s-robotics-ai-developers/ , site: https://humynlabs.ai/

The business: robots that do household and factory tasks need huge amounts of footage of humans doing those tasks. Humyn Labs built its own two-camera rig, records people doing their real work and chores (not staged demos) in 20+ countries with Brazil, India and South Korea as hubs, then uses software to track hands and object interactions and label every clip. Customers are frontier AI labs, robotics companies and world-model builders; 12 are paying, and about 90% of revenue is from the US.

The numbers are unusually concrete for a nine-month-old company: $10M revenue from April to October, $20M more in signed contracts, another $20M in deals not yet final. The debt has a plain explanation too. Humyn pays collectors daily or weekly, but AI lab customers pay in 30 to 60 days, so the faster it grows, the bigger the cash gap.

That working-capital line is the real story of physical-AI data. It looks like a software business to the labs buying it and like a labor business to the company running it. India's large workforce makes it cheap, and the same model has already drawn government scrutiny over in-home collection, so consent and supervision will matter as much as camera quality.
← Previous
LanceDB Raises $57.5M Series B for the Missing Data Layer in AI Training
← Back to all articles

Comments

Loading...
>_