Physical AI Goes Beyond Robots: Insights from an Nvidia Veteran Turned VC

Amulya Vishwanath spent seven years shaping NVIDIA’s robotics product strategy before leaving to found Techable VC. She argues that today’s robotics are still in a “Lego era,” where core components like actuators, grippers, software, and safety standards are being rebuilt from scratch.

Investment Focus Areas

Vishwanath defines physical AI as any system that interacts with the physical world—a molecule, a material, a medicine, or a machine on a factory line. She calls this “conscious automation,” emphasizing solutions that relieve pain points such as labor shortages in family‑run factories or burnout among surgeons.

Drawing on her Nvidia network, she treats her venture capital as a moat for founders, using deep relationships for due diligence and customer introductions. She notes the United States excels in early‑stage innovation and ecosystem support, while China offers scale and rapid deployment for manufacturing‑focused physical AI.

Why it matters for GPU / AI infrastructure: Training models that understand material properties, simulate robotic motions, or optimize factory workflows demands massive parallel compute. GPU clouds like AiGpu provide the scalable acceleration needed to turn physical‑AI concepts into deployable reality.