Former Nvidia robotics executive turned VC says field still maturing

By Splitfeed EditorialLast updated Oct 16, 2026, 16:40 UTC

On October 16, 2026, Amulya Vishwanath, a former Nvidia product marketing lead who spent seven years working on the company's robotics platforms before moving into venture capital, shared her current read on the robotics industry. She described the field as being in a "Lego era," a phrase meant to convey that the fundamental components, including actuators, robotic hands, software stacks, and safety standards, are still being designed and assembled rather than snapped together into finished products. She also argued that physical AI is a broader category than robotics alone, though the specific domains she has in mind were not detailed.

The significance of her position lies in the direct challenge it poses to a competing narrative that has been gaining momentum in technology and investment circles. That narrative holds that large foundation models are about to trigger a ChatGPT-style inflection point for robotics, meaning rapid, widespread capability gains driven by general-purpose AI trained at scale. Vishwanath's counter-argument, grounded in years of working inside one of the companies most central to that vision, is that the hardware and standards layer is not ready to support such a leap. The genuine dispute here is about timing and sequencing: whether software advances can pull physical robotics forward even when the underlying components remain immature, or whether the physical stack has to mature first.

Because no specific reactions have been published from the Pro-AI, Anti-AI, or Middle Ground camps yet, what follows reflects what each would typically argue about a story like this. Pro-AI voices would likely treat Vishwanath's framing as overly conservative, pointing to recent advances in manipulation, locomotion, and foundation models for robotics as evidence that the inflection point is closer than she suggests. Anti-AI voices would probably welcome her caution as a corrective to hype, using it to argue that investment capital is outrunning actual technical readiness and that premature deployment of unfinished systems carries real safety risks. The Middle Ground camp would most likely see her Lego-era metaphor as an accurate description of where things stand, treating it not as pessimism but as a reminder that foundational infrastructure work and software ambition need to proceed in parallel rather than the latter racing ahead of the former.

The clearest signals to watch will come from robotics hardware startups reaching production milestones, any announcements of general-purpose foundation models being deployed in commercial robotic settings at scale, and whatever investment theses Vishwanath's new venture practice makes public, since her actual portfolio choices will say more than any single interview about where she thinks value will be created during the foundational phase she describes.