Humyn Labs Raises $47M to Fuel Robotics AI with Real-World Training Data

Indian startup Humyn Labs has raised $47 million in Series A funding to provide real-world video data critical for training AI and robotics systems. The round includes $26 million in equity backed by major venture investors, and $21 million in venture debt designed to help with cash flow needs as the company scales, especially in light of extended payment cycles from clients. Revenue between April and October hit $10 million, with another $20 million in contracts signed and roughly $20 million more still under negotiation.

What Humyn Labs Does Differently

Humyn Labs specializes in capturing unposed video footage of people performing everyday tasks using its proprietary two-camera hardware. Beyond recording, its software meticulously annotates each frame—tracking hand movement, interactions with objects, and other detailed actions. The data collection spans more than 20 countries, with operations centered in India, Brazil, and South Korea. Local partners identify collection sites, supervise participants, and ensure real-world contexts rather than staged scenarios.

Backers, Market Reach, and Ethics

The equity portion of the recent round was led by Accel, joined by Cactus Partners and others including Prosus Ventures and Bajaj Alts AI Opportunities Fund. Venture debt came via Lighthouse Canton Growth Debt Fund II and Trifecta Capital. Among Humyn Labs’ twelve paying customers are robotics and AI labs focused on advancing world models; approximately 90% of Humyn’s revenue originates from the U.S.

Privacy has become a growing concern in the industry as companies collect data inside homes. To address this, Humyn Labs works through local partners to recruit people going about their normal activities—household chores, routine tasks—not acting out curated scenes. Participants are compensated, supervised, and closely overseen to ensure quality.

Leadership, Growth, and Scale

Founded in February by Manish Agarwal and Ishank Gupta—alumni of Nazara Technologies and AB InBev, respectively—Humyn Labs builds on their experience with KGeN, their prior contribution-based studio network in emerging markets. Currently, the team includes 51 full-time employees based largely in Bengaluru, supplemented by local contributors spread across collection regions.

Daily or weekly payments are made to those collecting the data, even as Humyn waits 30 to 60 days for full payments from clients. The venture debt is being used to bridge that liquidity gap without resorting to more equity dilution.

Overall, India is increasingly becoming a hub for training-data startups aiming to serve the global AI and robotics sector. The country’s large populations and emerging-market networks make it possible to gather diverse, real-life data at scale, though navigating consent and privacy remains crucial.

What this means. As robots and AI systems lean more heavily on real-world data to perform everyday tasks—from aiding in homes to autonomous operations—having access to richly annotated, diverse footage becomes essential. Humyn Labs is stepping into that gap, positioning itself as a bridge between emerging market realities and high-end robotics labs. Going forward, effectiveness in ethical data collection, quality annotation, and maintaining diverse geographical reach will likely determine who leads this emerging class of training data providers.