Tech Desk: India is emerging as an unexpected global powerhouse in the race to build advanced humanoid robots, not through software, but by exporting a unique commodity: human motion data. As robots transition from labs to real-world environments, the critical bottleneck isn’t computing power, but a massive shortage of real-world data showing how humans perform physical tasks.

US-based Instawork Robotics Lab (IRL) is betting big on India, launching its Instacore system here. This wearable five-camera setup captures synchronized video, hand movements, and spatial information as workers carry out their daily jobs. Instawork emphasizes that India’s allure isn’t just cheaper labor, but the unparalleled diversity of tasks and environments, from manufacturing to hospitality, crucial for training versatile physical AI models.

Unlike software AI, which can learn from vast digital datasets, physical AI demands an understanding of human interaction with the physical world—how objects are handled, how tasks vary. The demand is immense; Instawork estimates the industry needs millions of hours of data, a fraction of which has been collected so far.

This initiative presents a new economic opportunity, compensating businesses and workers for data collection, and potentially creating new employment categories in physical AI deployment. With engineering teams in India and the US, Instawork’s Bengaluru operation is also becoming a hub for hardware deployment and maintenance. Ultimately, India’s diverse workforce is poised to teach machines how the physical world truly operates.