Ex-Nvidia-turned-VC Amulya Vishwanath: ‘Physical AI is in its LEGO era’

Amulya Vishwanath, founder and managing partner of Techable Ventures, a venture firm focused on physical AI, told TechCrunch's Equity podcast that the robotics industry is in a "LEGO era." Before starting the firm she spent nearly seven years as a product marketer at Nvidia.

Her argument is that the foundational pieces, hands, actuators and core software, are still being rethought, and that those parts have to mature before they snap together into the breakthrough products everyone is waiting for. She also named talent as a decisive factor, with so much of it flowing to large companies like Anthropic and Nvidia.

We agree with the framing, and we wish more of the physical AI conversation sounded like it. The phrase "ChatGPT moment for robotics" has been repeated so often it implies the leap is imminent. A builder-blocks view is more grounded: robots fail in the physical world in ways chatbots never have to, and better components and software foundations are prerequisites, not details.

It is worth being clear about what this story is. It is a short interview with an investor whose firm benefits from enthusiasm about physical AI. There are no new numbers, deals or products here. That does not make the view wrong, but it is perspective, not evidence.

Her historical comparison, to the internet era and the Industrial Revolution, is optimistic, and we share some of that optimism. Robots that can handle dangerous, repetitive or physically punishing work would be a real good. But those earlier transitions also brought long stretches of displacement before broad gains. If the pieces do click together, the labor effects will land unevenly, and that deserves planning now, while the industry is still in its component phase.

The talent point is the sharpest part. If the best engineers keep consolidating at a few giants, the building blocks may end up owned by the same handful of companies that dominate AI software. Startups and the VCs backing them pitch themselves as the counterweight. Whether that holds is a real question.

Where this leads if it scales: a stack of interchangeable robot parts and shared software that lets many companies build capable machines. That is a healthier future than one closed platform. We are watching who ends up owning the standard pieces.

Before Nvidia, Vishwanath held a product marketing role at a company later acquired by Intel and worked as a digital signal processing engineer at Interconnect Systems, later bought by Molex. The interview did not disclose fund size or specific portfolio companies.