Back to Insights

AI Robotics Is Still in Its “Lego Era,” Investor Says

5 min read

Amulya Vishwanath, a former Nvidia product marketing leader for robotics platforms who is now an investor, says robotics is still in its “Lego era.” In a TechCrunch interview, she described core pieces—including actuators, robotic hands, software, and safety standards—as building blocks that are still being rebuilt.

That is a useful counterweight to the idea that one powerful AI model will suddenly make robots broadly ready for work. Vishwanath’s view is not a product announcement or a prediction that robotics will never become useful; it is an informed assessment that the underlying pieces are not yet settled.

Why Business Owners Should Care

Business owners are seeing plenty of demonstrations of AI agents, humanoid robots, warehouse systems, and machines that appear able to understand their surroundings. It is reasonable to ask whether a robot could help with fulfillment, food preparation, inspections, cleaning, stocking, or repetitive production work.

But a polished demonstration is not the same as a dependable operation. A business needs a system that can work safely in its actual environment, handle exceptions, fit into current processes, and be supported when something goes wrong. In physical settings, small variations matter: a different package shape, a wet floor, a crowded aisle, an employee working nearby, or a task that requires a delicate grip.

Vishwanath’s “Lego era” description points to that practical gap. Robotics depends on more than AI reasoning. Hardware has to move reliably. Software needs to interpret the environment. Safety practices need to be clear. The pieces also have to work together in a way a business can maintain.

For a small or medium business, that is less a reason to ignore robotics than a reason to separate near-term improvements from long-term bets. The best AI project is usually not the flashiest one. It is the one tied to a real bottleneck, with a manageable cost, clear ownership, and a way to measure whether it helped.

That same principle applies to information-based AI. For example, many teams can gain value sooner by improving meeting notes, call summaries, and searchable records with AI-powered transcription for business workflows. Those projects avoid the physical safety and hardware integration challenges that come with robots.

What to Watch Before Investing in Robotics

Start with the task, not the machine

If you are evaluating robotics, define the job in plain language before speaking with vendors. How many times is the task performed each week? How variable is it? What errors occur now? Does it require judgment, fine motor control, movement around people, or frequent changes to materials?

Robotics is more likely to be worth exploring where work is repetitive and controlled. A stable process with predictable objects and clear safety boundaries is fundamentally different from a changing, customer-facing environment. This is an operational assessment, not a promise that any specific robot will fit your business.

Ask how exceptions and safety are handled

A vendor demonstration should lead to practical questions: What happens when the robot cannot complete a task? Who can stop it? Who maintains it? What training is required? What conditions fall outside its intended use?

Do not treat safety standards as a technical detail someone else will solve later. Vishwanath specifically identified safety standards as one of the building blocks still being rebuilt. Before a pilot, involve the people who manage the workspace and the employees who will work alongside the system.

Build AI capability where it is ready now

Many businesses have valuable work that does not require physical automation: sorting incoming requests, finding answers in internal documents, drafting routine communications, or reducing manual data entry. Those use cases can teach your team how to set goals, review outputs, protect information, and measure results.

A structured AI readiness audit for your business can help identify those opportunities before you commit budget to a complex technology category. It can also reveal whether a process needs cleanup before automation, which is often the more valuable first step.

A Better Way to Think About “Physical AI”

The phrase “physical AI” can make it sound as though robotics is one category with one readiness level. In reality, a simple, purpose-built machine in a controlled setting may be much closer to practical use than a general-purpose robot expected to perform many changing tasks.

That distinction matters when reviewing proposals. Ask vendors to explain the exact task, operating conditions, support model, and limits of the system they are offering. Avoid filling in the missing details with assumptions based on videos, headlines, or claims about general AI progress.

The news here is not that small businesses should rush into robotics. It is that an experienced voice from the field is urging realism about how much foundational work remains. That perspective can help owners avoid spending money on an impressive-looking solution that does not match their day-to-day operation.

The PAD Take

Do not buy a robot because competitors are talking about AI. First, document one costly, repetitive workflow and identify whether the real issue is physical labor, poor process design, or information bottlenecks.

If robotics appears relevant, run a narrowly defined pilot with success measures for safety, uptime, employee effort, and cost. For most businesses today, start with proven AI workflows and build the operating discipline that will make future automation investments easier to evaluate.


More from the blog

Enjoyed this? Get more like it.

Practical AI insights, delivered monthly. No spam.

Unsubscribe anytime. We respect your inbox.

All Insights