Powered by Smartsupp

Humyn Labs Raises $47M Series A to Fuel Real-World Robot Training Data for U.S. AI Companies



By admin | Nov 18, 2026 | 3 min read


Humyn Labs Raises $47M Series A to Fuel Real-World Robot Training Data for U.S. AI Companies

India's Humyn Labs has secured $47 million in Series A funding as global demand surges for real-world data to train robots, positioning the company to help U.S. AI and robotics firms develop models capable of functioning in physical environments. The financing package includes $26 million in equity led by Accel, with backing from Cactus Partners, Prosus Ventures, BeeGlobal, Bajaj Finance's Finserv Intelligence and Bajaj Alts AI Opportunities Fund, Dallas Venture Capital, Lighthouse Canton, and Jump Trading. An additional $21 million comes as venture debt led by Lighthouse Canton Growth Debt Fund II, with Trifecta Capital participating.

Robotics companies worldwide are racing to build machines that can handle everyday human tasks, yet acquiring sufficient high-quality training data remains a significant hurdle. Startups such as Mecka AI, XDOF, and MicroAGI are already working on this challenge by gathering real-world data for robotics training, and Bengaluru-based Humyn Labs is aiming to compete in this space.

Humyn Labs has created proprietary two-camera hardware that captures footage of people carrying out routine tasks. The company's software then processes and labels this footage—tagging what occurs in each clip so AI systems can learn from it. According to co-founder Manish Agarwal, the startup collects data across more than 20 countries in Latin America and Asia, with Brazil, India, and South Korea acting as primary hubs.

Agarwal, formerly CEO of Indian gaming company Nazara Technologies, established Humyn Labs in February alongside Ishank Gupta, a former executive at brewing giant AB InBev. The pair had previously founded Kratos Studios, which is now known as KGeN.

EMBED_PLACEHOLDER_0

Recording human activity is only one piece of the puzzle, Agarwal noted. He explained that Humyn Labs also employs software to track hand movements, object interactions, and other details, transforming raw recordings into usable training data for AI models. "Scale, diversity, and quality are important aspects, and they need to work in tandem," Agarwal said.

The company reports having 12 paying customers, including frontier AI labs, robotics companies, and developers of world models. Gupta said approximately 90% of revenue originates from the U.S., though he declined to identify specific clients.

This growth has heightened Humyn Labs' working capital requirements, Agarwal explained, which accounts for the $21 million venture debt portion. The startup compensates workers and data collection partners on a daily or weekly basis but typically waits 30 to 60 days for payments from its AI lab customers. Borrowing helps bridge this gap and is less dilutive than issuing additional equity, he said.

In recent months, India has become a hub for collecting training data for what is termed physical AI, with startups like Human Archive and BuildAI leveraging the country's large workforce to record people performing everyday activities. However, this practice has prompted concerns about privacy and consent, especially when individuals are filmed inside their homes.

Agarwal stated that Humyn Labs gathers data from people engaged in their normal work or household chores rather than asking them to perform staged activities. The startup collaborates with local partners to identify collection sites and supervise the process. "The worker is a worker who is incentivized, but he needs to be monitored and supervised for the right input," he stated.

Humyn Labs presently employs 51 full-time staff, most located in Bengaluru, along with a network of local partners across the countries where it collects data, according to Agarwal.




Comments

Please log in to leave a comment.

No comments yet. Be the first to comment!