deniz.in

Markets

Weather

Loading weather

· via TechCrunch

Humyn Labs raises $47M Series A to supply real-world training data to U.S. robotics companies

Humyn Labs has raised $47 million in Series A financing ($26 million in equity led by Accel, $21 million in venture debt) to help U.S. AI and robotics companies train models that operate in the physical world.

Humyn Labs raises $47M Series A to supply real-world training data to U.S. robotics companies

Humyn Labs, a Bengaluru-based startup that records everyday human activity and converts it into training data, has raised $47 million in Series A financing to serve U.S. AI and robotics developers building models that can operate in the physical world.

According to TechCrunch, which reported the round on November 18, the financing combines $26 million in equity led by Accel with $21 million in venture debt. The equity tranche drew participation from Prosus Ventures, BeeGlobal, Bajaj Finance's Finserv Intelligence, the Bajaj Alts AI Opportunities Fund, Dallas Venture Capital, Lighthouse Canton and Jump Trading. The debt was led by Lighthouse Canton Growth Debt Fund II, with Trifecta Capital joining.

Turning chores into training data

Humyn Labs builds its own two-camera hardware for filming people as they carry out routine tasks. The footage is only half the product: the startup's software labels what happens in each clip, tracking hand movements, object interactions and other fine-grained detail so machine-learning systems can learn from it. Co-founder Manish Agarwal described recording as just one part of the challenge, with processing and annotation the other.

Collection spans more than 20 countries across Latin America and Asia, and its three main hubs are Brazil, India and South Korea, Agarwal said in an interview. Rather than staging activities for the camera, workers are filmed doing their usual work and household chores, with local partners identifying collection sites and supervising the process. "The worker is a worker who is incentivized, but he needs to be monitored and supervised for the right input," Agarwal said.

A gaming veteran's second act

Agarwal, the former CEO of Indian gaming company Nazara Technologies, founded the startup in February alongside Ishank Gupta, a former executive at brewing giant AB InBev. The pair previously co-founded Kratos Studios, which now operates as KGeN and runs a contributor network across emerging markets — a network Humyn Labs uses to arrange data collection in people's homes. The company currently has 51 full-time employees, most based in Bengaluru, on top of its web of local partners.

Growth explains the debt

Humyn Labs says it has 12 paying customers, a mix of frontier AI labs, robotics companies and developers of world models. Roughly 90% of its revenue comes from the U.S., Gupta said, though the company declined to name specific clients. Revenue reached $10 million between April and October, with another $20 million in signed contracts and a further $20 million in deals not yet finalized, Agarwal told TechCrunch.

That growth explains the unusual debt component of the round. The startup pays workers and data-collection partners daily or weekly, while its AI-lab customers typically take 30 to 60 days to settle their invoices. Debt covers that timing mismatch while diluting founders less than another equity sale would, Agarwal explained.

A crowded race, and a contested practice

The round lands amid intensifying competition to supply the real-world data that so-called physical AI systems need. Startups including Mecka AI, XDOF and MicroAGI are already collecting similar footage, and India has become a hub for the practice, with companies such as Human Archive and BuildAI recruiting Indian workers to film everyday tasks.

The model has also drawn scrutiny over privacy and consent, particularly when people are recorded inside their own homes. Agarwal argued that quality is the differentiator, noting that "Scale, diversity, and quality are important aspects, and they need to work in tandem."

Why it matters

The raise is a sign that the bottleneck for physical AI is shifting from model architecture to data supply. Robotics developers can no longer train capable systems from simulations or lab footage alone; they need large volumes of ordinary humans doing ordinary things, and that demand is being routed to lower-cost collection networks across emerging markets. The structure of the round tells its own story: a debt-heavy financing for a business that pays its workers weekly and waits months for frontier labs to pay their invoices. As this collection economy scales, the unresolved questions around consent and in-home recording are only getting sharper.

  • #robotics
  • #training-data
  • #physical-ai
  • #funding
  • #series-a
  • #india

Related posts