Nvidia (NVDA) is on the move...

It closed Friday at a record high and just broke above a $5 trillion market cap again.

Congrats to those followed my guidance and bought Nvidia in 2018. It’s up 3,000% since my initial recommendation.

And it’s not done yet.

That’s because Nvidia is having its “AWS moment”...

In other words, a big opportunity is forming behind the scenes. One investors can’t ignore.

Let me explain...

• Back in the early 2000s, Amazon (AMZN) was fast becoming America’s #1 online retailer.

But running a fast-growing e-commerce business created a huge technical headache.

As more people shopped on Amazon, it had to keep expanding the invisible machinery behind its website. That required more computing power, bigger servers, and more backend systems to keep the website fast and reliable—even during surges in traffic.

Before long, Amazon realized plenty of other businesses had the exact same problem. Startups, retailers, and software companies all needed computing infrastructure. But few wanted to build it from scratch.

So Amazon started renting out its extra capacity. That system became Amazon Web Services, or AWS.

At first, it looked like a small side project... nothing compared to Amazon’s retail empire. Today AWS generates most of Amazon's profits, despite making up only about a sixth of total revenue. The "side project" became the most profitable business Amazon has ever built.

• Nvidia is having its AWS moment today.

Most investors only think of Nvidia in terms of the artificial intelligence (AI) data center boom. But similar to Amazon, another opportunity is forming in the background...

I’m talking about robotaxis.

Robotaxis are one of those innovations that happened gradually, then suddenly.

For years, they were confined to tightly mapped test areas with long waitlists, limited hours of operation, and only a small number of vehicles on the road. Self-driving demos have existed since the 1980s. Carnegie Mellon ran one in 1986.

Today, they’re everywhere. Waymo has a fleet of 3,000 robotaxis. They clock in half a million rides each week across 11 cities.  And the company plans to expand to 20 more cities this year.

But that’s just one example...

Tesla (TSLA) is expanding fast, too. It only entered the market last year. Now, it’s already in three cities, with plans to launch in five more by the end of 2026.

Tesla also plans to start selling its "Cybercab" robotaxi with no steering wheel or pedals this year.

At $30,000 per vehicle, it’s pretty affordable. I don’t drive and I’d get one for sure.

• Here’s why this all benefits Nvidia.

A robotaxi is more than just a car with some fancy sensors and software. It’s a full AI system. And Nvidia sits at every layer of it.

Before a vehicle ever picks up a passenger, the self-driving system must learn how to handle the chaos of the real world, from bad weather, reckless drivers, strange road signs, emergency vehicles, sudden lane changes, unpredictable pedestrians. All the messy stuff that makes city driving hard.

Nvidia's GPUs provide the computing muscle to train these self-driving models.

But even a massive fleet can't encounter every possible danger in real-world driving. So developers build those scenarios in a virtual world instead.

Nvidia's DRIVE Sim platform is the digital proving ground. It lets engineers replay crashes, run edge cases, and simulate tricky road situations again to see how the software responds.

And once the software is ready, every vehicle still needs a powerful in-car computer.

Nvidia's DRIVE AGX system is the car's central nervous system. It combines inputs from cameras, radar, LiDAR, and ultrasonic sensors to build a live picture of the road. Then it decides when to brake, steer, change lanes, or avoid danger, in milliseconds.

Train the model. Simulate the model. Run the model in the car.

From the first training run to live traffic, Nvidia powers every step. And the more robotaxis hit the road, the more money Nvidia makes from every layer of the stack.

• Uber (UBER) recently revealed a plan to launch 100,000 robotaxis.

Uber won’t build the cars, of course. It’ll partner with other carmakers. The key is that they’ll be powered by Nvidia’s stack.

Lyft (LYFT), Grab Holdings Ltd. (GRAB), and Bolt are all using Nvidia’s DRIVE Sim for their self-driving initiatives, too… as are Mercedes-Benz, Nissan, BYD, and Hyundai. And the list goes on.

Just like Amazon competitors in the 2000s couldn't build their own AWS, so they ended up paying Amazon for it, every other automaker on Earth needs a self-driving stack and almost none of them can build it from scratch.

Designing the chips, training the models, and running the simulation infrastructure costs tens of billions of dollars. Tesla can spend that. Mercedes can't. Hyundai can't. Uber can’t.

So they all rent it from the company that already built it.

And similar to back then, these deals aren’t making headlines yet... and they’re not bringing in lots of money compared to Nvidia’s other businesses. But they’re proof of just how big the robotaxi megatrend could become for Nvidia.

AWS eventually became the backbone of cloud computing. Today, Nvidia is turning into the backbone of robotaxis.

Stephen McBride

Chief Analyst, RiskHedge

PS: Nvidia’s robotaxi push is a perfect example of how the biggest investing opportunities often start quietly. I cover these early-stage shifts and what they mean for investors in my free letter, The Jolt. If you’d like to stay ahead of these kinds of disruptions, you can join us here.