Amulya Vishwanath, who spent seven years leading product marketing for NVIDIA’s robotics platforms, has launched Techable VC — her new early-stage firm focused on physical AI. She’s writing first checks in robotics, materials science, and life sciences. Not software-only AI. Not chatbots. Machines that move, sense, and interact with the real world, but not yet at scale, and not yet unified by a single model. She calls the current state the “Lego era”: actuators, robotic hands, safety standards, and foundational software are still being rebuilt, piece by piece. That’s the headline. Not a launch. Not a product. A diagnosis. From someone who helped sell NVIDIA’s robotics stack for nearly a decade. And it contradicts the bullish narrative pushed by others, like General Intuition CEO Pim de Witte, who told TechCrunch’s Equity podcast earlier this year that a single general foundation model is imminent, one that would make today’s specialized robotics work redundant. That moment, de Witte said, would be robotics’ “ChatGPT moment.” Vishwanath disagrees. She says it hasn’t arrived. Not even close. Her fund isn’t betting on a sudden inflection. It’s betting on the slow, uneven work of assembling reliable hardware, certifying systems for real environments, and aligning AI with physical constraints. Heat, torque, latency, failure modes. The stakes aren’t theoretical. Robotics startups face higher capital intensity, longer regulatory paths, and tighter margins than their software peers. And investors are noticing: CoreWeave shares steadied Thursday after a sharp slump tied to concerns over debt-financed AI infrastructure build-outs, a reminder that physical layers cost more, scale slower, and break differently. Techable VC’s first investments haven’t been named. No dollar figures, no portfolio companies, no launch dates. Just a thesis. And timing. Vishwanath is one of the first women executives to spin out of NVIDIA and found her own VC firm. Her pitch isn’t about speed or virality. It’s about durability. Physical AI won’t go viral. It’ll get certified. Tested. Revised. Deployed in labs, factories, hospitals, then, maybe, homes. The rollout starts there. Not in an app store.