Nvidia announced during its GTC conference on Monday, March 16, that Uber will begin rolling out a fleet of Level 4 autonomous vehicles in Los Angeles and San Francisco in 2027, as part of both companies’ broader self-driving efforts, expanding on an intent to deploy 100,000 vehicles running on Nvidia’s Drive Hyperion platform that was announced during GTC in Washington in October 2025.
Level 4 autonomy refers to vehicles capable of operating without human intervention within defined geographic and operational parameters, a classification that represents the commercial threshold at which ride-hailing economics begin to change materially because the cost of a human driver, which accounts for the majority of per-trip expense in the current Uber model, is removed from the equation.
The Los Angeles and San Francisco markets are among the most heavily contested in the autonomous vehicle sector, with Waymo already operating commercially in both cities and Tesla’s Robotaxi service preparing its own expansion, making Uber‘s 2027 entry a direct confrontation with the most technically advanced operators already on the road.
Nvidia’s Drive Hyperion platform provides the sensor and compute infrastructure for the autonomous system, with Nvidia positioning itself as the neutral supplier to multiple autonomous vehicle operators rather than a direct competitor, a posture consistent with the company’s broader strategy of being the infrastructure layer beneath AI applications rather than the application itself.
The commercial timeline of 2027 places the rollout in the same window as Vera Rubin’s broad availability and Groq 3 LPU shipments, suggesting that Nvidia’s automotive ambitions are being deliberately aligned with the generation of chips that will power the AI inference required for real-world autonomous decision-making at meaningful scale.
Uber’s involvement carries its own strategic logic: the company has historically struggled with the capital intensity of building autonomous technology itself, having sold its Advanced Technologies Group to Aurora in 2020, and partnering with a platform that Nvidia is supplying to multiple operators gives Uber exposure to autonomous economics without the proprietary technology development costs that burned through capital across the industry’s first development cycle.
The 100,000-vehicle figure attached to the partnership covers both Los Angeles and San Francisco as initial deployments, suggesting a national expansion roadmap that would follow once the initial commercial operations demonstrate safety and unit economics at meaningful scale.
For Nvidia’s stock, autonomous vehicles represent a long-duration revenue category that complements rather than substitutes for the near-term AI data centre demand that currently dominates the company’s financial profile, giving investors a second growth vector to incorporate into multi-year models as the current infrastructure super-cycle inevitably begins to mature.