Bengaluru-based Humyn Labs has secured $47 million in Series A financing to provide real-world training data for U.S. artificial intelligence and robotics developers, according to a report published by TechCrunch on November 18, 2026. The funding package includes $26 million in equity and $21 million in venture debt, positioning the startup to compete in the growing market for physical AI data collection. As companies worldwide work to build robots that can handle everyday human activities, obtaining sufficient quality data to train these systems remains a significant hurdle.
The equity portion of the financing was led by Accel, with backing from Prosus Ventures, BeeGlobal, Bajaj Finance's Finserv Intelligence and Bajaj Alts AI Opportunities Fund, Dallas Venture Capital, Lighthouse Canton, and Jump Trading. Lighthouse Canton led the $21 million venture debt component, joined by Trifecta Capital. The startup currently serves 12 paying clients, including frontier AI labs, robotics firms, and world model developers, with roughly 90% of its revenue originating from the United States. Between April and October, Humyn Labs generated $10 million in revenue and has locked in another $20 million through signed agreements, with a further $20 million in potential deals still under negotiation. The company gathers data across more than 20 countries throughout Latin America and Asia, with Brazil, India, and South Korea functioning as its primary collection centers.
Co-founder Manish Agarwal, formerly CEO of Indian gaming company Nazara Technologies, told TechCrunch that the startup has built proprietary two-camera hardware to film individuals performing daily tasks, then applies its software to process and label the recordings by tagging activities in each clip so AI systems can learn from them. Agarwal launched Humyn Labs in February alongside co-founder Ishank Gupta, a former AB InBev executive, leveraging their earlier venture KGeN's network of contributors across emerging markets to facilitate data collection inside people's homes. "Scale, diversity, and quality are important aspects, and they need to work in tandem," Agarwal stated in the interview. The startup employs 51 full-time staff members, mostly based in Bengaluru, plus a network of local partners throughout its data collection territories.
The report notes that Humyn Labs captures footage of people conducting their regular work or household activities rather than staging scripted scenarios, with software tracking hand movements, object interactions, and other particulars to convert raw video into usable training data. Agarwal explained that the $21 million debt component addresses the startup's working capital requirements, since Humyn Labs compensates workers and data collection partners daily or weekly but typically waits 30 to 60 days to receive payment from its AI lab customers. India has become a center for collecting training data for physical AI in recent months, with startups like Human Archive and BuildAI also tapping the country's large workforce to record everyday human tasks, though the practice has sparked concerns about privacy and consent when filming occurs inside homes. The company works with local partners to identify collection sites and supervise the process, with Agarwal noting that workers are incentivized but require monitoring to ensure proper input quality. For robotics and AI developers racing to create systems capable of operating in real-world settings, access to diverse, high-quality footage of human behavior across different cultures and environments may determine which models succeed in physical tasks. The venture debt strategy reflects a broader calculation about maintaining equity ownership while funding rapid expansion in a capital-intensive business where payment timing creates cash flow gaps.

