NVIDIA's ambitions in autonomous driving: From ground to sky, the next growth engine of the all-powerful chip emperor

According to the Zhitong Finance APP, at the GTC conference held in Washington, NVIDIA (NVDA.US) unveiled a series of "super hardcore" new products and roadmaps covering seven major areas including AI, quantum computing, open-source models, American manufacturing, enterprise computing, robotics, and 6G, causing its stock price to soar by about 5%, reaching a new historical high with a total market value approaching $5 trillion. However, NVIDIA's strategic layout in the field of autonomous driving is particularly noteworthy in this "all-round show": from planning to deploy 100,000 autonomous taxis with Uber (UBER.US), to partnering with Lucid (LCID.US) to bring L4-level capabilities into high-end passenger vehicles, and planning to mount IGX Thor on Joby Aviation's (JOBY.US) eVTOL. Let’s break down this "autonomous driving chess game" from the road to the sky.

100,000 Robotaxi Computing Base: Uber Hands Over the Key to Autonomous Driving Scale to NVIDIA

NVIDIA has once again placed the "calculations" of autonomous driving at the forefront of the mobility platform. In January of this year, the company began deep technical collaboration with Uber—Uber provided some operational data to help NVIDIA refine its AI models and chips; within less than a year, the two parties presented a more ambitious implementation plan at GTC Washington: Uber plans to deploy up to 100,000 Robotaxis based on Drive AGX Hyperion starting in 2027, with plans to have Stellantis, Foxconn, and other potential production partners manufacture in batches starting in 2028, with an initial delivery volume of at least 5,000 vehicles, initially focusing on the U.S. market before expanding globally.

For NVIDIA, this not only means gaining a major customer but also signifies that its chips, sensor suites, and software toolchains will be embedded in the daily operations of a mobility giant, creating a continuous and replicable revenue stream.

Moreover, the scale effect is the core leverage of NVIDIA's layout this time. Uber has clarified that the 100,000 target includes 20,000 Lucid Gravity and Nuro models finalized in July, while the remaining gap will be opened to more than a dozen existing or potential partners—Avride, May Mobility, Momenta, Pony.ai, and others can supply based on the same Drive platform. This way, NVIDIA can rapidly expand its chip shipment scale without additional investment in scenario construction.

At the same time, Uber has committed to establishing a "robot taxi data factory," which will return 3 million hours of exclusive driving data within three years, with NVIDIA providing processors, AI models, and data filtering and simulation tools to complete the full-link closed loop from labeling, scene mining to synthetic data generation. The faster the algorithm iteration, the shorter the chip upgrade cycle, and NVIDIA's bargaining power in the L4-level autonomous driving market will be more solid.

What’s even more noteworthy is the extension of the commercial path. Uber will be responsible for the end-to-end fleet management of these vehicles, including remote assistance, charging, cleaning, maintenance, and customer service; while NVIDIA, through a unified hardware platform, will obtain high-frequency, multi-scenario return data and provide downstream manufacturers with a "ready-to-go" mass production sample. For NVIDIA, once the fleet of 100,000 vehicles is realized as planned, it will have the first "mobility platform-level" L4 validation case

From Luxury Cars to Autonomous Vehicles: Can Lucid Leverage NVIDIA DRIVE AV to Bring L4 to High-End Passenger Cars?

NVIDIA adds another ally, and the gap in high-end autonomous passenger vehicles is expected to close. Lucid Group announced on Tuesday that it will develop an L4 autonomous driving solution based on the NVIDIA DRIVE AV platform, with the initial project targeting the current Gravity SUV, first enhancing driving assistance features to a higher level; then transitioning to NVIDIA's next-generation DRIVE AV full stack, relying on a fusion architecture of cameras, radar, and lidar to achieve "driverless" L4 capabilities.

This collaboration reflects NVIDIA's gradual validation strategy of "from assisted driving to fully autonomous driving," enabling technological iteration and data accumulation on the Lucid Gravity SUV, ultimately completing L4 capability development using the DRIVE AV platform and verifying the feasibility of its hardware integration in passenger vehicles. This process not only promotes the commercialization of autonomous driving technology but also provides crucial support for NVIDIA's technological layout in the luxury car market.

Lucid's interim CEO Marc Winterhoff did not provide a timeline for implementation but clearly stated that "bringing this technology to consumers" is a priority for the company. Lucid indicated that future model updates could access the DRIVE AV system via OTA, and the current Air and Gravity models could also connect to the system through OTA in future updates.

Additionally, Lucid is adopting a dual-track strategy of "consumer passenger cars + autonomous taxis": collaborating with Uber and Nuro to develop a fleet of 20,000 Gravity SUV autonomous taxis while internally developing L4 technology. Through this collaboration, NVIDIA verifies the practical integration capabilities of its hardware platform in the passenger vehicle sector, covering both consumer markets and mobility service scenarios.

Overall, by leveraging Uber for autonomous taxis and delivery fleets, and Lucid for consumer passenger vehicles, NVIDIA achieves full-scenario coverage from B-end to C-end. Meanwhile, Uber's data resources and Lucid's vehicle platform complement each other, allowing NVIDIA to reduce single-path risks through a multi-partner strategy, accelerating technological iteration and commercialization processes.

Bringing Autonomous Driving to the Skies: Joby Aviation Plans to Install IGX Thor, Allowing eVTOL Air Commuting to Utilize NVIDIA's Computing Power

It is worth mentioning that NVIDIA has also moved the "chessboard" of autonomous driving from the road to the sky, with its latest move being a partnership with Joby Aviation. According to public information, Joby is collaborating with NVIDIA to advance the development of the company's autonomous flight technology "Superpilot" by integrating NVIDIA's IGX Thor computing platform. Supported by NVIDIA's Blackwell architecture, IGX Thor aims to power the next generation of physical artificial intelligence applications.

By combining IGX Thor with Joby Aviation's aircraft design and flight testing capabilities, NVIDIA and Joby Aviation can accelerate the development of autonomous aviation technology. By integrating advanced computing capabilities, the aircraft will be able to autonomously determine, request, and follow optimal flight routes, adapt to weather conditions, air traffic control instructions, or emergencies, and even make judgments with human-like intuition In addition, the platform will enable aircraft to predict when system components need maintenance and alert the crew before component failures occur.

After the announcement of the collaboration, Joby Aviation's stock price rose more than 10% in after-hours trading. Thus, NVIDIA has extended the application scenarios of its autonomous driving hardware from ground vehicles to urban air mobility, adding a category of aviation use cases to its autonomous driving product line