Nvidia’s NEW AI Self Driving Cars

NVIDIA'S NEW AI SELF-DRIVING CAR So Jefferson Huang just walked out on stage at GCC 2026 in San Jose today and said 7 words that should make every person on the planet stop and pay attention. The chat GPT moment for self-driving cars has arrived. Not it's coming, not it's close, not we're working on it has arrived and then he backed up and backed it up with something that will completely change how you think about getting in a car for the rest of your life. NVIDIA ANNOUNCED 4 BRAND NEW CAR PARTNERS FOR THEIR ROBOT TAXI READY PLATFORM, BYD, HYUNDAI, NISSAN, GEELY, THAT'S NOT 4 RANDOM SMALL COMPANIES, BYD IS THE LARGEST ELECTRIC VEHICLE MAKER ON THE PLANET EARTH, HYUNDAI SELLS MILLIONS OF CARS A YEAR, NISSAN HAS BEEN ON THE ROAD SINCE 1933, GEELY OWNS VOLVO, OWNS LOTUS, OWNS POLESTAR, THESE ARE NOT START UPS, THESE ARE THE BIGGEST NAMES IN THE CAR BUSINESS SIGNING ON TO THE SAME AI SYSTEM AND THOSE 4 NEW NAMES JOIN A LIST THAT ALREADY HAS GM, TOYOTA, MERCEDES, BENZ, JAGUAR, LANDROVER, VOLVO, RIVIAN, XIAOMI AND ALL SORTS ON BOARD, STOP AND COUNT THAT, THAT'S MANY AND LITERALLY AT THIS POINT THEY'VE ACTUALLY GOT 19 IN TOTAL OF THE WORLD'S BIGGEST CAR BRANDS, ALL BUILDING ON THE SAME AI BRAIN FROM NVIDIA BUT HERE'S THE THING THAT SHOULD MAKE YOUR JAW DROP, THIS ISN'T JUST ABOUT CARS THAT DRIVE THEMSELVES, NVIDIA AND UBER ARE PLANNING AND PARTNERING TO LAUNCH A GLOBAL FLEET OF FULLY AUTONOMOUS ROBOTAXIS STARTING IN LOS ANGELES AND SAN FRANCISCO IN THE FIRST HALF OF 2027 AND THEN SCALING ACROSS 28 CITIES GLOBALLY BY 2028, NOT 2040, NOT SOME FAR OFF SCI-FI FUTURE, 2027, that's just 12 months away, right?

And if you think this is hype, just wait until I show you what's already happening on the streets right now because the race is already underway and most people have absolutely no idea how far along this actually is. And I keep saying this, the biggest technology shift in human history is not happening in some lab, it's happening on the 101 freeway, it's happening in downtown San Francisco, it's happening in Phoenix, Arizona, it's happening right now whilst you're sitting here watching this. Let me show you exactly what's going on and why it matters and changes everything. Let's start with a number I don't think enough people have heard.

Waymo, the robo-taxi company owned by Google's parent company Alphabet, began 2025 doing about 175,000 rides per week. They ended 2025 doing about 455,000 rides per week. That's a 157% increase in 12 months, 157% growth in one year. In 2025 alone, Waymo more than tripled their annual volume to 15 million rides, surpassing 20 million lifetime rides.

Let me put that into plain English. Waymo gave more robot cab rides in the single year of 2025 than they gave in their entire history up to that point, combined. And they are just one company, they're not even the biggest player anymore in terms of global ambitions. They served over 14 million trips in 2025 alone and they're on a path to hit 1 million fully autonomous rides every week by the end of 2026.

1 million rides a week from robot cars with no driver. If you told someone this in 2025, they would have laughed at you. In 2025, it's the goal and the path is already mapped out. Sorry, 2026.

Now, here's where NVIDIA comes in and why today's announcement is so much bigger than most people understand. Think about what just happened. Waymo spent years building their own AI from scratch. They spent billions of dollars.

They mapped every street in San Francisco. They hired thousands of engineers. They tested for over a decade and they got to 450,000 rides a week, which is impressive. But what Jason Huang announced today at GCC 2026 is a completely different model.

NVIDIA is not building one robo-taxi fleet. They're building the operating system that every car company on the planet uses to build their robo-taxi fleet. Think about what Android did to smartphones. Before Android, every phone company had to build their own software.

Samsung had their own system. LG had their own system. Motorola had their own system and it was slow, it was expensive, and it was messy. Then Google released Android, one system that any phone maker could build on.

And suddenly a thousand different phone companies can make smartphones without reinventing the wheel from scratch. And NVIDIA is doing exactly the same thing for self-driving cars. NVIDIA's Drive Hyperion platform is designed to be the standard architecture for level 4 autonomy. And at GCC 2026, BYD, Geely, Nissan, and Hyundai all announced they're adopting it.

Alongside a growing list that already includes Stellantis, Lucid, and Mercedes Benz. One platform with 19 car companies. Hundreds of millions of cars on the road around the world. And right at the center of all of it is a piece of AI technology that NVIDIA just updated.

They call it Alpameya. And I need to explain what it does because it's the actual brain that makes this whole thing work. Here's how Jason Huang described it at CES earlier this year. Not only does it take sensor input and activate steering wheel brakes and acceleration, it also reasons about what action it's about to take.

Reasons and what action it's about to take. This is a massive difference from how self-driving cars used to work. I want to make sure you understand this because it's the key to why this moment is different from all the moments before it. Old self-driving systems worked on pattern matching.

The car's brains saw a stop sign a million times during training. So when it sees a stop sign in real life, it knows to stop. It's like a really fast lookup table. See this thing?

Do that thing. The problem is the world is infinite. The world does not follow a script. You've got a stop sign in the middle of a four-way intersection but it's a police officer standing in the middle waving cars through.

What does the car do? The old system has never seen a police officer waving you past a stop sign. It doesn't know what to do. It freezes and it probably makes the wrong call.

Alpameya 1 is a 10 billion parameter model that allows a self-driving car to think more like a human so it can solve complex edge cases like how to navigate a traffic light outage and a busy intersection without previous experience. Without previous experience. It can reason through situations it's never seen before. It thinks through the problem the way a human driver would.

Step by step cause and effect. Unlike traditional systems that just detect objects and plan a path, Alpameya uses chain of thought reasoning. It processes video input and generates a trajectory but crucially it also outputs the logic behind its decision. The logic behind this decision.

You know what that means for regulators, for courts, for the insurance companies that have to figure out who's liable when a robo-taxi gets in an accident. It means there's a paper trail. There's a reason the car can explain what it was thinking and why it made the call it made. That's not a small thing.

That's a massive step toward the legal and regulatory framework that's required to make commercial robo-taxis possible at that scale. At GCC 2026, NVIDIA also unveiled Alpameya 1.5 which takes driving video, motion history, navigation data, and natural language instructions as inputs and outputs driving trajectories with traceable reasoning. Developers can now steer driving behavior directly through text prompts and since launch the Alpameya portfolio has been downloaded by more than 100,000 developers. 100,000 developers building on this platform right now.

That is the Android moment. That is the moment where the ecosystem starts to build itself and you literally can't stop it. Now let me bring in a skeptic voice here because I know what some of you are thinking. You're thinking we've heard this all before.

We've heard and been hearing this for 10 years. In 2016 Elon Musk said there would be fully autonomous Tesla cars driving cross-country by 2018. In 2019 he said there would be a million robo-taxis on the road in 2020. In 2021 people were writing breathless articles about self-driving being a year away and every single year it didn't happen and I get that.

That's a completely fair and legitimate reaction but here's the difference between then and now and this difference is crucial. Those promises were about technology that didn't exist yet. They were promises about future that had to be invented from scratch. What Jensen Huang announced today is not a promise about technology that has to be invented.

It's a deployment announcement about technology that's already running. Waymo is already doing 450,000 rides a week, not in a test environment, not with a safety driver, real rides, real passengers and real money changing hands. In December of last year NVIDIA offered reporters and analysts an hour-long ride through San Francisco in a 2026 Mercedes-Benz CLA sedan. The safety driver behind the wheel said the car was driving itself for 90% of the ride.

That's not a demo, that's production testing. That's a car that's going to be on the road and in your Uber app within 18 months. The skepticism was correct in 2018, the math was not there, the data was not, the compute was not there but in 2026 all three are there. Here's the historical arc that I need you to understand because this is how big technological shifts always work.

First someone proves the concept works. Waymo did that in Phoenix starting in 2020. Small fleet, controlled environment, fully autonomous, no driver works right. The second the technology gets dramatically cheaper and more capable.

That's what NVIDIA's chips have done over the last four years. NVIDIA's automotive revenue jumps 69% year over year and CEO Jensen Huang has said the auto business is on its way to eventually being a trillion dollar business. The economics of running this stuff are changing fast. Third, a platform emerges that lets the whole industry build on one set of tools instead of reinventing everything from scratch.

That's what NVIDIA's hyperdrive Hyperion is. That's what Alpameo is. That's exactly what happened today. And fourth, the deployments start compounding.

One city becomes five cities. Five cities becomes 20 cities. 20 cities becomes 100. Waymo currently offers paid driverless rides in 10 major cities and has done over 200 million autonomous miles on public roads.

200 million miles. 200 million real world miles driven by AI with no human at the wheel. That's not a prototype. That's not a proof of concept.

That's an operating reality. And now NVIDIA just handed the keys to 19 car companies and said here's a system, here's an AI brain, here's a safety sack, here's the simulation tools. Let's go build the next 10 million rides together. Here's the specific deployment plan that NVIDIA and Uber announced today.

Uber will begin scaling its global autonomous fleet starting in 2027 targeting 100,000 vehicles supported by a joint AI data factory built on the NVIDIA Cosmos platform. 100,000 vehicles globally started in 2027. Central to this deployment is NVIDIA's Alpameo, a next generation reasoning-based AI model designed to handle complex scenarios like unpredictable construction zones or erratic pedestrian behavior using chain of thought logic. Think about what that means for NVIDIA, sorry, Uber as a business.

Uber's entire cost structure is built around paying human drivers. That's their biggest expense. Every single fare split, every driver, every incentive, every surge pricing negotiation, every driver support ticket, all of it exists because humans are in the cars. Take the human out of the car and Uber's economics become something completely different.

Uber CEO Dara said autonomous technology holds enormous promise to make transportation safer, more reliable, more accessible. By expanding our partnership with NVIDIA and combining advanced AI with Uber's global network and operating experience we are laying the foundation for an increasingly multiplayer AV world. Multiplayer, that's a keyword. Uber isn't betting on one self-driving company winning, they're building a marketplace.

Way more rides on Uber in some cities. NVIDIA powered cars on Uber in 28 cities. Aurora powered trucks on Uber Freight. Uber becomes the app player on top of every autonomous vehicle company on the planet.

That's not a taxi company, that's an AI powered logistics operating system. Let me stack some examples here so you feel the weight of this. BYD sold 1.76 million electric vehicles in 2024. They just signed on to NVIDIA's self-driving platform.

That is not a boutique experiment. When BYD ships self-driving cars they will ship them at a scale that most American companies can't even imagine. Hyundai. They just announced an expanded partnership with NVIDIA that covers everything from level 2 plus to full level 4 robo-taxi design.

Hyundai will combine its software-defined vehicle capabilities and large-scale fleet data with NVIDIA's AI computing infrastructure and explore cooperation with its autonomous driving joint venture Motional to accelerate innovation in next-generation autonomous mobility services. Motional, that's the joint venture Hyundai built specifically for robo-taxis. It's already operating in Las Vegas and now they're combining it with NVIDIA's full stack. Mercedes-Benz.

NVIDIA's CFO said on the earnings call, we are now in production with our full stack solution for Mercedes-Benz. Starting with the new CLI hitting roads in the next few months. In production right now as you're watching this. The 2026 Mercedes-Benz CLA, a car you can actually buy, runs NVIDIA's AI driving system.

Not in a test, not in a demo, in a car at a dealership. And this is just the beginning because here's the number that should make you sit up straight. NVIDIA's automotive revenue for the full year of 2026 hit a record 2.3 billion dollars, up 39 percent, driven by continued adoption of their self-driving platforms. 2.3 billion dollars in one year from automotive alone.

And Jensen Wang said that this path and this business is on a path to being worth trillions, not billions, trillions. Now here's where I want to make a personal connection to what this actually means for you. Not just as someone watching the car industry from a distance, but someone who either drives for a living, owns a business that depends on transportation, or just gets in a car every single day. There are about 3.5 million professional truck drivers in the United States.

About 1.5 million taxi and ride share drivers. Millions more in delivery, logistics, transportation, adjacent jobs around the world. The standard reassuring line you should hear from economists is, don't worry, new technology always creates more jobs than it destroys. And historically that's been true.

The cotton gin didn't end employment. The printing press didn't end employment. The internet definitely didn't end employment. But here's the thing, the speed matters.

When technology disruption happened slowly over generations, people had time to adapt. New industries emerged. The kids who grew up after the cotton gin didn't train to be cotton hand pickers. They trained to be textile machine operators, then factory supervisors, then logistics managers.

The question today is whether this transition is happening slow enough for the workforce to adapt. And honestly, I don't know the answer to that. I don't think anyone does. What I do know is that pretending it isn't happening is the worst possible strategy.

If you're driving for Uber today, or running a logistics business, or managing a fleet, the single most useful thing you can do right now is understand exactly where this technology is, how fast it's moving, and what skills are going to be valuable on the other side. Because the Uber driver who gets ahead of this isn't the one who fights it, it's the one who becomes the fleet manager for a company operating 500 autonomous vehicles. The logistics manager who learns how to optimize the self-driving truck network. The entrepreneur who figures out which businesses become more valuable when transportation costs drop by 60%.

It's not hypothetical, it's a real opportunity right now, and it's available for people who are paying attention. Here's something I want to address directly, because I've had this conversation with a lot of people. There's a version of this story that gets told on the internet that goes like this. AI is going to take all the jobs, the robots are coming, we're all going to be poor and unemployed, and the billionaires are going to own everything.

And there's an opposite version of this that goes, don't worry, technology always worked out, progress is always good, trust the process. I think both of these stories are too simple. The honest version is this, massive technology shifts create massive opportunities and massive disruption. They happen to both happen at the same time, in the same city, it's affecting different people in different ways.

The people who capture the opportunity are not always the ones who deserve it the most, and the disruption doesn't fall evenly. That tension is real and it deserves to be taken seriously. But the worst response to that tension is to be uninformed, because uninformed people don't shape outcomes, they just experience them. If you understand what NVIDIA is building, you can make better decisions about what skills to develop, what businesses to start, what investments to make, what conversations to have with your kids about what they should study as well.

That's why I keep doing these videos, not to hype AI, not to scare you, but to make sure you have the real picture of what's actually happening, backed by real numbers, real companies, and real deployments. Because the chat GPT moment for self-driving cars, it didn't just arrive today, it started arriving a long time ago, just most people weren't watching. Let's talk about the competitive landscape here, because I think it helps you understand just how locked in this NVIDIA story is. Right now there are three main players in the self-driving race from a technology standpoint.

You've got Waymo, owned by Alphabet, backed by 16 billion dollars in fresh capital, at a 126 billion dollar valuation as of February 2026. The current operational leader, over 200 million autonomous miles on public roads, paid driverless rides in 10 major cities. They are the gold standard for what fully autonomous looks like today. You've got Tesla, taking a completely different approach, using cameras only, building a massive fleet of cars, already on the road that feed data, back to train a neural network.

Elon's bear is at fleet size, and data volume beats everything else in the long run. The jury is still out on whether that's right. And then you've got NVIDIA, right, not building cars or even the robo-taxi fleet directly, but building the AI platform that every other car company and robo-taxi operator is building on. Here's the thing that makes NVIDIA's position fascinating.

They win regardless. It doesn't matter which other player wins, right? If Waymo dominates robo-taxis, Waymo uses NVIDIA chips to train the AI. If Tesla wins with a different approach, no problem.

Tesla uses NVIDIA chips in their supercomputers to train FSD. If BYD floods the market with cheap self-driving cars built on drive Hyperion, well that's NVIDIA revenue. If Hyundai, Nissan, Mercedes, Lucid, Stellantis all ship cars with NVIDIA's full-stack system, that's NVIDIA's revenue, right? It's the same business model that made NVIDIA the most valuable chip company in the history of the world.

Sell the shovels to everyone digging for gold, and you win no matter who finds the gold. Goldman Sachs projects 35,000 robo-taxis on US roads by 2030. A market analyst forecasts 40% annual compound growth in the autonomous vehicle sector and compound annual growth every single year for the next four years. That's not a niche market, that is an industry transforming from curiosity into infrastructure.

Now, I want to pause for a second here and talk to you directly because here's something I've been thinking a lot about, right? Most of the people watching this video are not car industry insiders. Most of you are creators, entrepreneurs, marketers, professionals, people running businesses or building careers who are trying to understand this AI moment and figure out what it means for your life. And I keep coming back to one thing.

Every major technology shift creates a window. A window where the people who understand what's happening early can position themselves before the crowd arrives, before it gets priced in, before everyone already knows. That window is almost always shorter than most people think. The people who made the most from the internet didn't wait until 2005 to start a website.

The people who made the most from mobile apps didn't wait until 2020 to build for smartphones. The people who got ahead of the AI moment didn't wait until it was obvious. The self-driving moment is not fully obvious yet. Most people still think of it as a future thing, as something that might happen someday.

We've been over the numbers today, right? 450,000 Waymo rides a week, 19 car brands on NVIDIA's platform, robotaxis in LA and San Francisco in 2027, 28 cities by 2028. This is not a future thing. This is a now thing.

And the businesses, the skills and knowledge that are going to be valuable in the world, those are being built right now by people who are paying attention. That's exactly why I built the AI Profitable because the gap between the people on this technology and the people who don't is becoming an economic gap, a real gap, a gap that shows up in who gets the contracts, who builds the products to win, who gets hired for the highest value roles and who makes money from those shifts instead of losing their job to them. If you want to be on the right side of that gap, the AI Profitable is where I work through that stuff every day. Real strategies, real implementations, real business applications of AI that you can start using today.

Not theory, not hype, actual tools and workflows that my community members are using right now to get ahead. Link in the comments description or just go to the aiprofitable.com and come join us. Let me go back to the technology story because there's something about NVIDIA's strategy that I haven't explained yet and it's important. So with self-driving cars, they face what engineers call the long tail problem.

And once you understand it, you understand why this technology took us as long as it did, right? Think about driving. 99% of your daily driving routine is just routine, right? You stay in your lane, you stop at red lights, you yield at intersections, you go the speed limit.

If you trained an AI on 99% of driving situations, it would be a pretty good driver. But 1% of driving is weird, right? A mattress falls off a truck on the highway. A police officer directs you to drive the wrong way on a one-way street because of an emergency, right?

A traffic light is out and five cars from four directions all arrive at the intersection at the same time. These are rare situations, but they're not that rare. If you drive every day for a year, you'll probably encounter dozens of situations that don't fit the standard patterns. And an AI that gets confused by the situations, well that's no bueno, right?

Could slam on the brakes at 60 miles an hour or do something dangerously unpredictable. And this is why the old approach to self-driving failed, right? You can't collect enough real-world data to cover every weird situation. And you know, there are many of them.

The tail is too long. NVIDIA's key strategic insight at GCC 2026 is that they want to turn the robotics data problem into a compute problem. Instead of needing more real-world data, they use simulation pipelines and synthetic data generation to replace expensive real-world data collection, making raw compute power not flea-sized, see bottleneck, for training better models. Read that again.

Instead of needing more real-world data, they use simulation. NVIDIA's Cosmos platform can generate synthetic driving data, realistic simulations of weird edge case driving scenarios, traffic light outages and rainstorms, mattresses and highways, all sorts of crazy stuff like that. In simulated environments that look and feel like the real world, run at massive scale, all powered by NVIDIA's own chip. So instead of needing to drive 10 million real miles to collect the data for a weird scenario, you just simulate it a billion times if you want to, in a virtual world that teaches AI exactly what to do.

This is why NVIDIA's automotive revenue has grown 39% in a year. This is why 19 car brands have signed on to their platform. Because NVIDIA doesn't just make the chip that goes in the car, they made the simulation environment you train the AI in. They made the safety certification stack that proves AI is safe to regulators.

They made the operating system the car runs on. They made the AI model that does the reasoning. It's a complete end-to-end system. And if you're a car company that doesn't have $10 billion in 15 years to build all of that from scratch, you sign on to NVIDIA's platform and you get it all in one.

This is why everyone's signing up to it. Now at GTC 2026, NVIDIA also introduced Halos OS, a three-tier safety architecture based on a certified safety standard that includes what's called an NCAP five-star safety stack. NCAP is the same safety rating we use to evaluate how safe cars are for humans. Five stars, that's the standard a car has to meet before a country says it's safe to sell to regular people.

NVIDIA is not waiting for regular regulators to figure out how to certify autonomous vehicles. They're getting there ahead of the regulators. They're building a certification framework in. They're making safety a feature of the platform, not an afterthought.

And that's huge because the number one thing that slowed robo-taxi deployments is not the technology. The technology has been good enough for years in controlled environments. The slow part is the regulatory approval process, proving to governments that the AI is safe enough to put real passengers in with no backup. So let me bring this home with the bigger story underneath it.

There's a race happening, not just between car companies, not just between Waymo and Tesla and NVIDIA. It's bigger than that. NVIDIA's automotive platform is already adopted by automakers, tier one suppliers, startups, and we've already seen a lot of big companies sign up to it already. Now Jensen Huang said at GCC, if we have the technology to solve self-driving cars, self-driving trucks and robots, this will be the largest technology industry the world has ever seen.

The largest world has ever seen. That's a bigger claim than the internet. It's a bigger claim than mobile computing. It's a bigger claim than the industrial revolution.

And sitting here in March 2026, with 450,000 robot rides happening every week in real cities, with 19 car brands on one AI platform, with robo-taxis launching in LA and San Francisco in 12 months and 28 cities by 2028, with Alfa Romeo reasoning through complex driving scenarios the way a human does, I'm not saying he's right. I'm saying you can't dismiss it. The Chad Chippity moment that Jensen Huang is talking about is real. It's a moment where suddenly the technology goes from impressive demo to obvious utility.

The moment where people stop asking whether it works and start asking how to use it. Chad Chippity had that moment in November 2022. A hundred million users in two months. Not because the technology suddenly improved, because suddenly it was packaged in a way where everyone could see that it worked and understand how and what to do with it.

Self-driving is having that moment. Maybe not as dramatically overnight, but across 28 cities, across 19 car brands, across a hundred thousand developers building on NVIDIA's platform, the shift from demo to utility is underway. I want to leave you with one more thing before we close. I talked to a lot of people who were sitting on the fence about learning this stuff, about really committing to understanding AI and what it means for their life.

And the thing I hear most often is this, I'm not sure this is as big as people say. And I understand that, right? There's been a lot of hype. There's been a lot of promises that didn't pan out.

Cynicism is a reasonable response to years of over-promising. But here's the thing that I keep coming back to. The downside of taking this seriously and learning everything you can about AI and automation, and then it turns out to be slightly less transformative than people said, is that you learn some really, really powerful and useful skills. The downside of it, deciding it's all hype, waiting to see what happens, and then it turns out the optimists were right, is that you're on the wrong side, the biggest economic shift in a generation, right?

And those two downsides are not equal. The cost of being informed is time. The cost of being uninformed could be your entire livelihood. That's not a threat, that's just maths.

Waymo is doing 450,000 driverless rides per week. NVIDIA just signed 19 car brands, right? This is not maybe, this is a timeline. And the question you ask yourself is whether you're prepared for it.

I think you are. That's why you're here. Keep watching, keep learning, keep asking the hard questions. If you want to go deeper, if you want to be in a community of people who are actively building, thinking, and adapting in real time as this stuff develops, come find me in the AI Profit Building.

Link in the comments description or go to theaiprofitbuilding.com. The car driving you home in 2027 might not have a human in the front seat. The question is whether you built something valuable in the world that the car is moving through. Get to work.