Former Nvidia Marketer Says Physical AI Is Still in Its “LEGO Era”

Physical AI has not yet reached its breakout moment, according to venture capitalist Amulya Vishwanath, who says the robotics industry is still assembling the basic pieces needed for a true leap forward. Speaking to TechCrunch’s Equity podcast, the former Nvidia product marketer argued that the field is in a “LEGO era” rather than waiting on a single ChatGPT-style breakthrough.

That distinction matters because it frames how investors, founders, and customers should think about robotics today: less as a finished revolution and more as a period of infrastructure-building, talent competition, and rapid iteration on core hardware and software components.

Why Vishwanath says robotics is still early

Vishwanath, now founder and managing partner of Techable Ventures, focuses her firm on physical AI — the broad category that includes robotics systems, embodied intelligence, and the software layers that let machines interact with the real world. Her view is that the industry is not yet at the moment when a single product or model will unlock everything at once.

Instead, she sees the market as a construction zone. Teams are still developing the essential components that robots need to function reliably, affordably, and at scale. Those components include robotic hands, actuators, sensing systems, control software, and other foundational parts that can later be combined into more capable machines.

Vishwanath said the sector has been waiting for a “ChatGPT moment” for robotics, but described the reality as a “LEGO era,” where foundational blocks are still being redesigned and assembled.

Her point is that progress in robotics is likely to arrive in layers. First come the building blocks, then the integration of those parts, and only later the kind of mainstream adoption that changes how industries operate. In her view, that sequence is similar to other major technology cycles, including the early internet and the industrial transformation that preceded it.

What does “the LEGO era” mean for physical AI?

It means the sector is still defining its standards, its core components, and its winning product categories. The metaphor suggests a market where innovation depends on combining many smaller advances rather than discovering one all-purpose solution overnight.

In practical terms, that could mean a startup builds a better robotic gripper, while another focuses on motion planning, and another develops a software stack that lets machines adapt to variable environments. None of those pieces may feel revolutionary on their own, but together they can form the basis of a scalable robotics platform.

This is one reason physical AI is attracting attention from founders and investors even before a mass-market robot emerges. The companies that define the underlying architecture of the sector may become the equivalents of picks-and-shovels suppliers in a broader industrial boom.

How the field differs from consumer AI

Unlike software-only AI products, physical AI must contend with the messy constraints of the real world. Robots need hardware that can survive wear and tear, respond to unpredictable settings, and perform tasks safely alongside people.

That raises the bar for deployment. A chatbot can be iterated quickly through software updates, but a robot often requires mechanical redesigns, field testing, and expensive integration work. Vishwanath’s analysis suggests those complications are exactly why the industry remains in its foundational phase.

• Robotics must solve both software and hardware challenges.
• Physical systems require testing in unpredictable environments.
• Adoption often depends on integration with existing industrial workflows.
• Scaling a robot business typically takes longer than scaling a pure software product.

Who is Amulya Vishwanath and why does her view matter?

Vishwanath’s perspective carries weight because she has spent years working close to the companies and technologies shaping modern AI infrastructure. Before launching Techable Ventures, she spent nearly seven years as a product marketer at Nvidia, a company central to the AI hardware boom.

Before Nvidia, she held a similar role at a company later acquired by Intel, and earlier worked as a digital signal processing engineer at Interconnect Systems, which was later bought by Molex. That mix of technical and go-to-market experience gives her a broad view of how deep-tech businesses evolve from early prototypes to commercial products.

Her career path also helps explain why she is positioning Techable Ventures as a specialist investor. Rather than backing general AI startups across the board, her firm is aimed specifically at physical AI opportunities, where technical complexity and talent access can be decisive advantages.

Why talent is becoming a competitive moat

Vishwanath also emphasized that talent is one of the most important factors shaping the U.S. robotics and physical AI market. In her view, strong teams can be the difference between a startup that merely experiments and one that becomes a category leader.

She said her firm sees talent access as one of its competitive advantages. Techable Ventures aims to help founders connect with the engineers, operators, and technical leaders they need to build quickly in a market where many of the best people are being pulled toward large companies.

Vishwanath said her network, built over years in the sector, makes it easier for founders to view her as an early call when they need help hiring or finding the right technical partners.

That issue is not trivial. In fast-moving AI markets, hiring can determine whether a startup ships a usable product before competitors do. Vishwanath pointed out that top talent is often drawn to major players such as Anthropic and Nvidia, intensifying pressure on early-stage founders to recruit aggressively and creatively.

Why startup teams matter so much in robotics

Robotics companies usually need expertise from multiple disciplines at once. Mechanical design, embedded systems, machine learning, supply chain management, product development, and field deployment all have to work together.

That creates a premium on experienced hires who can operate across boundaries. For investors like Vishwanath, backing a promising founder is only part of the job; helping that founder assemble a credible team can be just as important.

1. Technical co-founders help build the product foundation.
2. Experienced operators help move prototypes into production.
3. Recruiting networks help startups access scarce expertise faster.

How the robotics market may evolve from here

Vishwanath expects the current phase of robotics development to continue producing a stream of foundational advances before the market reaches a more obvious inflection point. That means investors and founders may need patience as the sector works through its technical stack.

Rather than looking for one dramatic breakthrough, she suggests watching for steady improvements in the underlying layers that make robots more useful and more commercially viable. Once those layers mature, more ambitious products may become easier to build and deploy.

The broader pattern is familiar in technology history. New industries often spend years defining core standards before they become widely adopted. Vishwanath’s comparison to the internet era and the Industrial Revolution reflects that long-view approach: the biggest shifts often look gradual until they suddenly do not.

What investors are watching now

In physical AI, investors are paying close attention to several signals that can indicate momentum:

• Better hardware components that are cheaper and more reliable.
• Software frameworks that simplify robot training and control.
• Clear business cases in logistics, manufacturing, and other industrial settings.
• Teams that can combine AI expertise with real-world deployment experience.

Those signals matter because robotics startups often need to prove more than software companies do. They have to show that their machines can work not just in demos, but in noisy, changing, and economically meaningful environments.

What makes physical AI different from the rest of the AI boom?

Physical AI is different because it bridges digital intelligence and the physical world. Instead of only generating text, images, or code, these systems must perform actions in real spaces — lifting, sorting, navigating, grasping, or assisting in industrial settings.

That makes the category broader and slower-moving than consumer AI, but also potentially more transformative. If robots become more adaptable and affordable, they could reshape factories, warehouses, labs, and other environments where repetitive labor or precision handling matters.

For now, Vishwanath’s message is that the industry should be judged by its building phase, not by unrealistic expectations of instant disruption. The important work, she argues, is happening below the surface: in the components, the software interfaces, and the talent pipelines that will support the next generation of machines.

Timeline: from deep tech to physical AI investor

The following timeline summarizes Vishwanath’s background and the current stage of the physical AI market she described.

Why the “ChatGPT moment” may take longer in robotics

A chatbot can suddenly feel revolutionary once it becomes easy to use at scale. Robotics, by contrast, must overcome physical limitations, safety concerns, and deployment friction before it can have a similarly visible public moment.

That does not mean progress is stalled. It means the path to a breakout is more cumulative. Each new actuator, sensor system, or control model can add capability, but the industry may not see a single universally recognized turning point until many of those innovations are working together in real products.

Vishwanath’s argument is ultimately one of timing. The robotics market, she suggests, is not late — it is early. The companies that endure through this stage may be the ones that help define the standards for the next one.

What founders should take from Vishwanath’s outlook

Founders building in physical AI may want to think less about hype and more about infrastructure. The most valuable companies in an emerging market are often those that solve a painful technical bottleneck or make the rest of the ecosystem easier to build.

They may also need to plan for a longer runway. Because robotics products require iteration in both software and hardware, teams often face slower commercial cycles than software-only startups. That makes capital, patience, and access to the right people especially important.

Vishwanath’s view suggests that the winners in physical AI will likely be defined by a combination of technical depth, execution speed, and the ability to recruit exceptional talent before competitors do.

In other words, the robots of the future are being assembled now — piece by piece, in a market still deciding what its first true standard will look like.