Venture capital is flowing into physical AI, yet a critical data gap remains unresolved. Travis Kalanick's industrial robotics venture Atoms, which targets sectors like food production, mining, and transportation, closed a $1.7 billion round led by Andreessen Horowitz in July. Skild AI tripled its valuation in seven months after raising $1.4 billion in January for an "omni-bodied" foundational model. However, physical interaction data cannot be scraped from the internet; every example must be performed, recorded, and annotated by hand, making data the defining constraint. Deccan AI, headquartered in the Bay Area with its largest operations base in Hyderabad, is addressing this by closing a $25 million Series A round in March, led by A91 Partners, to scale post-training data operations and expand into robotics data. The company builds post-training data, evaluations, and reinforcement learning environments, drawing on a contributor network of over a million people, with clients including Google DeepMind and Snowflake. According to Mordor Intelligence, the data annotation market crossed the $3 billion mark globally in 2026. Navaneeth Bodla, Head of AI at Deccan AI, describes physical AI as being at a GPT-2 moment, noting that scaling works but requires inverse scaling loss, where hardware and compute catch up. Deccan AI produces multimodal training data using egocentric video, multi-camera rigs, and motion capture setups with 23 cameras plus inertial sensors for millimetre