The Human Labor Behind India's Humanoid Robot Training Hub

In the city of Karur, Tamil Nadu, a new sector of the AI economy is emerging as local workers are hired to physically teach humanoid robots how to navigate the human world. This burgeoning data annotation industry is transforming mundane household chores into a critical pipeline for artificial intelligence development.

At the center of this shift is Objectways, a data annotation firm with offices in the U.S. and India. In its Karur labs, employees perform repetitive domestic activities, such as folding towels, while wearing head-mounted cameras like GoPros or smartphones. These workers operate within simulated environments, including model bathrooms, to capture the precise visual and spatial data required for robots to develop dexterity. To process this massive volume of footage, Objectways utilizes Amazon SageMaker, providing the resulting datasets to Fortune 500 multinationals.

The Shift to Embodied AI

India has long served as a global hub for digital data processing, but the industry is pivoting from text-based annotation to "embodied AI." While early AI training focused on virtual assistants and self-driving cars—areas where Objectways founder Ravi Rajalingam previously focused his efforts—the current frontier is physical interaction. This requires a "human-in-the-loop" process where AI does not just analyze data, but learns by observing human movement in three-dimensional space.

The Economic Paradox

This trend underscores a growing paradox in the AI revolution: as robots become more autonomous, they rely more heavily on an invisible human workforce. This has created a distinct economic divide where high-value intellectual property is developed in Western markets, while the physical labor of "teaching" the AI is outsourced to the Global South. The financial reality for these workers is stark; reports indicate they earn approximately 250 rupees per hour (roughly $3.00), with some sources citing rates as low as $2.62 per hour.

Market Implications

The scale of this effort is driven by massive commercial projections. The humanoid robot market is expected to reach approximately $38 billion by 2035, according to Goldman Sachs. As companies race to deploy robots in homes and warehouses, the demand for high-fidelity, human-demonstrated data will likely increase.

What remains to be seen is the long-term stability of these roles. While the industry is currently creating jobs in Karur, the very data these workers provide is intended to enable robots to perform the same tasks autonomously. The industry continues to expand, but the tension between creating current employment and training future replacements remains a central theme of the embodied AI boom.