Robotics is still stuck in the Lego era, not on the verge of a ChatGPT moment

Amulya Vishwanath spent seven years leading product marketing for Nvidia's robotics platforms. She says the field is still rebuilding its basic building blocks — actuators, robotic hands, software, safety standards. That view cuts against the prevailing narrative that a general foundation model will deliver a "ChatGPT moment" for robotics, mirroring the chatbot's late-2022 launch. General Intuition CEO Pim de Witte said earlier this year that a single such model would render much of today's specialized robotics work obsolete.

Vishwanath, who founded and runs Techable VC, defines Physical AI more broadly than robots: a molecule, a material, a drug, a machine on a production line. On TechCrunch's Equity podcast she explained that the best opportunities solve big, concrete problems — family factories that can't hire, surgeons burning out. She calls the approach "conscious automation," a term that stresses practical problem-solving over flashy demos.

Her Nvidia tenure built a network she now uses for due diligence and to connect founders with potential customers. Vishwanath wants entrepreneurs to see their investor as a moat, not just a source of capital. As one of the first women to spin out of Nvidia into an early-stage fund, she backs robotics, materials science and life sciences — sectors she says demand patience and a deep grasp of the value chain.

On global competition, she points to different advantages on each side: the U.S. leads in software innovation and foundation models, while China excels at manufacturing, supply chain and rapid deployment. Humanoids, she stresses, still need to prove utility beyond dancing and demos — a test the market has yet to pass.