The 'Lego Era' of Robotics
Amulya Vishwanath, who spent seven years leading product marketing for Nvidia's robotics platforms, argues that robotics remains stuck in what she calls its "Lego era" β a phase in which the fundamental building blocks of the field are still being assembled and rebuilt.
Those components include actuators, robotic hands, software infrastructure, and safety standards. Until these pieces mature and interoperate reliably, Vishwanath contends, the sector is not poised for the kind of breakthrough moment that generative AI experienced with ChatGPT in late 2022.
Her assessment cuts against a growing chorus of optimists who believe a single general-purpose foundation model will soon unlock robotics' "ChatGPT moment." Among them is Pim de Witte, CEO of General Intuition, who told TechCrunch's Equity podcast earlier this year that one general foundation model would render much of today's specialized robotics work redundant.
From Nvidia to First Checks
Vishwanath now serves as founder and managing partner of Techable VC, an early-stage firm investing in physical AI across robotics, materials science, and life sciences. On a recent episode of TechCrunch's Equity podcast, she spoke with Dominic-Madori Davis about where she is deploying her first checks and what she learned as one of the first women executives to spin out of Nvidia and launch her own fund.
What 'Physical AI' Meant Then vs. Now
The conversation traced the evolution of physical AI from 2018 β when the term carried a narrower, more hardware-centric meaning β to 2026, where it now spans foundation models, simulation, and cross-disciplinary applications. Vishwanath highlighted industrial and surgical use cases as areas moving well beyond the viral spectacle of dancing robots.
'Conscious Automation' and the Hype Cycle
Vishwanath introduced the concept of "conscious automation" β automation designed with deliberate awareness of its own boundaries and failure modes β as a framework for cutting through the sector's hype. Her "Lego era" metaphor serves a similar purpose: a reminder that compelling demos do not equal deployable systems.
Bottlenecks, Geopolitics, and Near-Term Promise
Other topics covered in the episode include:
- The biggest bottlenecks facing physical AI founders β from hardware supply chains to data scarcity
- The U.S. vs. China race in physical AI development and deployment
- Whether humanoids are destined for entertainment rather than practical labor
- The most promising near-term application areas for robotics investors
Vishwanath also made a case for why founders should treat their VC as a moat β not merely a source of capital, but a strategic partner in navigating the field's technical and regulatory complexity.
Listen to the Full Episode
The complete conversation is available on YouTube, Apple Podcasts, Overcast, Spotify, and all major podcast platforms. Follow Equity on X and Threads at @EquityPod.
Chapters:
00:00 Intro
01:03 Amulya's path from Intel to Nvidia, and why she left
04:39 What physical AI meant in 2018 vs. now
06:12 Beyond dancing robots: industrial and surgical use cases
09:31 What "conscious automation" means
12:12 Cutting through the hype: the "Lego era" of robotics
16:43 Why your VC should be your moat
20:30 Biggest bottlenecks for physical AI founders
22:30 U.S. vs. China in physical AI
25:43 Are humanoids just for entertainment?
28:10 The most promising near-term areas
29:22 Outro
via TechCrunch AI