Tesla (TSLA 2.28%) stock rocketed to a new all-time high in 2025, even though its electric vehicle (EV) sales saw their largest plunge in the company's history. Normally, that would be very strange, but investors are placing early bets on the future success of Tesla's upcoming product platforms, like the self-driving Cybercab robotaxi.
Some Wall Street firms, including Cathie Wood's Ark Investment Management, predict the Cybercab will create extremely valuable business opportunities in areas like autonomous ride-hailing, which could send Tesla stock soaring over the long term -- even from its present sky-high valuation.
However, the Cybercab already faces stiff competition, and it hasn't even hit the road yet. At the 2026 Consumer Electronics Show (CES) last week, Nvidia (NVDA 1.28%) unveiled a new family of artificial intelligence (AI) models that will transform its existing DRIVE platform, and potentially make autonomous vehicles a reality for almost any car company in the world. Here's why that's very bad news for Tesla.
Image source: Nvidia.
Nvidia's self-driving platform is already very popular
Around 90% of Nvidia's revenue comes from its data center business, where it sells industry-leading graphics processing units (GPUs) for AI development. But the company also has an automotive segment headlined by its DRIVE platform, which includes all of the hardware and software a car manufacturer needs to produce a self-driving vehicle.
DRIVE Hyperion is Nvidia's most powerful platform to date. It's specifically designed for Level 4 autonomy, which means vehicles can drive on their own in designated areas without any human input. It includes two AGX Thor in-vehicle computers, which are built on the company's flagship Blackwell chip architecture. It also includes an entire sensor suite with 14 cameras, 12 ultrasonics, nine radars, and one LIDAR system.
But Hyperion's software stack could be the real difference maker over the long term. It includes Nvidia's DriveOS, which not only governs the platform's autonomous driving capabilities, but also runs a series of AI-powered cockpit experiences. And then there is the Alpamayo family of open-source AI models, which is what caught the industry's attention at the CES conference last week.
NASDAQ: NVDA
Key Data Points
The Alpamayo ecosystem includes Nvidia's physical AI data set, which features over 300,000 real video clips filmed by cars driving in 2,500 cities around the world, in addition to the AlpaSim system, which can simulate real-world driving experiences. These tools give car manufacturers a foundation for building vehicles with true Level 4 autonomy, without having to spend years collecting data to train their own models.
In other words, Alpamayo makes the DRIVE platform more powerful than ever. Leading global car brands like Toyota, Mercedes-Benz, Jaguar, Land Rover, Volvo, Hyundai, and more, are already using it to accelerate their autonomous ambitions, but I would expect several more to join them.
Tesla's Cybercab still has a long road to commercialization
Tesla's EV sales declined by 8.5% to 1.63 million cars in 2025, as competition ramped up and eroded the company's market share in important markets like Europe. But CEO Elon Musk believes autonomous vehicles are the future of mobility, so he's more focused on bringing the Cybercab robotaxi to market than he is on reviving passenger EV sales.
Ark Invest believes the Cybercab could generate $756 billion in annual revenue for Tesla by 2029, mainly by providing autonomous ride-hailing services. For some perspective, the company generated less than $100 billion in total revenue during 2025 from all of its existing businesses, so Ark's forecast implies substantial growth.
However, the Cybercab isn't expected to enter mass production until the end of 2026, so it won't be generating meaningful revenue until mid-2027 at the earliest. Plus, Tesla's full self-driving (FSD) software -- which will power the Cybercab -- isn't approved for unsupervised use anywhere in the U.S. yet. If this isn't resolved in the next few months, the robotaxi might be grounded before it even hits the road. In my opinion, these headwinds put Ark's 2029 forecast in doubt.
NASDAQ: TSLA
Key Data Points
While Nvidia could help almost any car manufacturer pose a threat to Tesla in the robotaxi business in the future, I want to mention one company that is already miles ahead: Alphabet's (GOOG 1.40%)(GOOGL 1.47%) Waymo. It's already completing over 450,000 paid autonomous ride-hailing trips across five U.S. cities every single week, so the Cybercab will be playing catch-up from the moment it enters service.
Tesla's sky-high valuation opens the door to significant downside
Investors who are thinking about buying Tesla stock today because of the future potential of its robotaxi business need to consider its valuation. It's trading at a price-to-earnings (P/E) ratio of 297 as I write this, making it more expensive than any other company valued at $1 trillion or more -- and it's not even close.
In fact, Tesla is a whopping six times more expensive than Nvidia, so investors who want to ride the autonomous driving boom might want to consider looking there instead.
Data by YCharts.
Tesla stock is basically priced for perfection, meaning it could suffer significant downside if the robotaxi business hits any speed bumps on the road to commercialization. This is becoming more probable as competition ramps up, especially because Nvidia has a stellar track record of dominating new industries.
