Jensen Huang told the GTC 2026 audience that “the ChatGPT moment of self-driving cars has arrived,” backing the claim with concrete deployment commitments: Uber will launch NVIDIA Drive AV-powered robotaxi fleets across 28 cities on four continents by 2028, starting in Los Angeles. Four new automotive OEM partners — BYD, Hyundai, Nissan, and Geely — signed on to the NVIDIA autonomous driving stack.

The announcements, covered live by Yahoo Tech and CNBC, represent NVIDIA’s most aggressive push into autonomous vehicles since the company first launched its Drive platform.

The “ChatGPT Moment” Claim

Huang’s language choice is calculated. When ChatGPT launched in November 2022, it didn’t introduce new technology — transformer-based language models had existed for years. What changed was accessibility. Suddenly, 100 million people could interact with AI directly, and the resulting demand wave destroyed companies building inferior alternatives while minting new ones that could ride the platform.

Huang is claiming the same inflection point for autonomous vehicles. The underlying technology — lidar, computer vision, path planning — has been in development for over a decade. What’s changing, per NVIDIA’s framing, is that the compute stack has reached the point where deployment economics work at scale. Uber committing to 28 cities across four continents by 2028 is the kind of concrete timeline that either validates or falsifies the claim within two years.

The OEM Partner List Tells a Geopolitical Story

The four new partners — BYD, Hyundai, Nissan, and Geely — span three continents and both sides of the US-China divide. BYD is the world’s largest EV manufacturer by unit sales and Geely owns Volvo Cars and Polestar. Both are Chinese companies.

NVIDIA selling autonomous driving technology to Chinese automakers while the US government maintains export controls on advanced AI chips creates a tension that Huang did not address on stage. The Drive AV stack runs on NVIDIA silicon — the same category of high-end chips subject to Biden-era (and now continuing) export restrictions. Whether BYD and Geely can actually deploy NVIDIA-powered autonomous systems in China without running into export control barriers remains an open question.

Hyundai and Nissan fill out the Asian manufacturing base, with Hyundai’s Ioniq EV line and Nissan’s existing ProPILOT ADAS system providing natural integration points for NVIDIA’s stack.

Uber’s 28-City Bet

Uber’s commitment is the most commercially significant piece of the announcement. The company abandoned its own self-driving program in 2020 after selling its autonomous vehicle unit to Aurora. Partnering with NVIDIA to deploy robotaxis represents Uber’s re-entry into autonomous driving — but as a fleet operator running someone else’s technology rather than a technology developer.

Twenty-eight cities across four continents by 2028 is aggressive. For context, Waymo — widely considered the leader in deployed autonomous vehicles — currently operates commercial robotaxi service in four US cities (San Francisco, Los Angeles, Phoenix, and Austin). Uber is projecting seven times that footprint in two years, though the announcement did not specify how many vehicles per city or whether all 28 cities would have fully driverless (no safety driver) operations.

The Los Angeles launch city is strategic: LA is Uber’s largest US market by ride volume, and California’s regulatory framework for autonomous vehicles is the most developed in the country.

What the ChatGPT Analogy Actually Implies

If Huang’s analogy holds, the consequences extend beyond NVIDIA and Uber. The ChatGPT moment didn’t just create OpenAI’s business — it triggered a consolidation that killed dozens of AI startups building chatbots, coding assistants, and content tools that couldn’t compete with a platform-level deployment. Companies like Jasper, Copy.ai, and Character.ai saw their addressable markets shrink overnight.

The self-driving equivalent would be a consolidation of the fragmented autonomous vehicle industry around a small number of full-stack providers. Companies like Motional (Hyundai/Aptiv joint venture, which paused operations in 2024), TuSimple, and smaller autonomous trucking startups would face the same platform squeeze that ChatGPT inflicted on chatbot startups.

Huang’s choice to invoke ChatGPT at GTC — the same conference where he positioned OpenClaw as the foundational infrastructure for AI agents — frames autonomous vehicles as agentic AI deployed at physical scale. A robotaxi is an agent with a car. The software architecture problems (perception, planning, execution, safety constraints) map directly onto the agent framework challenges the rest of GTC is discussing.

Whether 2026 is actually the year self-driving crosses from perpetual promise to deployed reality will depend on execution, not keynote language. But Huang has now staked NVIDIA’s credibility on the timeline — and 28 cities by 2028 is specific enough to be proven right or wrong.