Nvidia’s founder and CEO Jensen Huang announced at the company’s annual developer conference on Tuesday plans to deploy its Level 4 autonomous driving software through partnerships with ride-hailing companies.
The plans include both Uber — with which Nvidia had already established a partnership — and the European ride-hailing company Bolt.
Both companies will operate fleets in Europe powered by Nvidia’s Drive AV software.
Uber x Nvidia
Nvidia detailed its ride-hailing partnership with Uber, announced late last year, revealing on Monday that it will cover across 28 cities in four continents by 2028.
The fleet powered by Nvidia‘s Drive AV software will be first deployed in Los Angeles and San Francisco next year.
The two companies will implement a phased deployment strategy in each launch city, starting with a fleet of data-collection vehicles to help train the Nvidia Alpamayo engine on city-specific driving nuances.
Following an initial operator-led phase, Uber’s fleet will gradually transition to fully driverless Level 4 operations.
This rollout is designed to scale robotaxi services to 28 cities across North America, Europe, Australia, and Asia by 2028.
Saying Uber is “building one of the world’s most expansive autonomous ride-hailing platforms,” Jensen Hung said its company will connect its “large ecosystem of robotaxi-ready partners to the Uber network to bring the magic of robotaxis to cities worldwide.”
Uber’s Autonomous Fleet
Last October, Nvidia and Uber announced they were partnering to expand autonomous robotaxi and delivery fleets using Nvidia‘s AI-powered Drive platform.
By then, the companies said Stellantis would deliver at least 5,000 Drive AV-equipped vehicles for Uber’s operations in the US and internationally, with Uber managing the fleets — including charging, maintenance, and customer support.
“Uber and Nvidia are working together to support a broad global Level 4 ecosystem — including Aurora, Avride, May Mobility, Momenta, Motional, Nuro, Pony.ai, Waabi, Wayve, and WeRide — to advance autonomy across passenger mobility, trucking, and delivery,” an official statement read.
Last July, Uber also partnered with Nuro and Lucid Motors to deploy a Robotaxi service using Lucid Gravity vehicles and Nuro’s software.
Lucid‘s interim CEO Marc Winterhoff said last week that all engineering testing vehicles had since been delivered and the program remains on track to debut in late 2026.
Additionally, Amazon subsidiary Zoox, which is developing purpose-built autonomous vehicles for robotaxi services, has struck a deal with Uber to have its vehicles available in the Uber app in Las Vegas this summer.
Uber aims to offer driverless rides across 15 cities by the end of 2026.
It already has driverless options in Atlanta, Austin, Dallas and Phoenix in the US, as well as several Middle East cities.
Bolt
In Europe specifically, Nvidia is also partnering with Bolt, a ride-hailing platform based in Tallinn, Estonia.
The companies will co-develop an AI-driven autonomous vehicle platform, leveraging Bolt’s extensive ride-hailing and car-sharing data.
The initiative will use Nvidia‘s Hyperion hardware, AGX Thor processors, and AI tools like Omniverse, Cosmos, and Alpamayo to build Level 4-ready robotaxis that meet Europe’s strict safety and GDPR standards.
The program prioritizes European data sovereignty, ensuring all fleet data complies with GDPR (EU’s General Data Protection Regulation) and stays under local control.
Additionally, open-source tools and frameworks will support local innovation and reduce dependence on non-EU technology ecosystems, a statement read.
Other ride-hailing platforms with which Nvidia is collaborating include Grab in Southeast Asia and Lyft in the US.
Partnering with OEMs
Earlier this year, Nvidia launched the Alpamayo-R1, which it calls “the world’s first thinking, reasoning, autonomous vehicle AI” at the Consumer Electronics Show (CES).
The visual-language-action (VLA) model is available open-source, as they aim to speed up the development of autonomous vehicles, allowing for any automaker to deploy a competitive AV platform.
By then, Jensen Huang pointed out that while Tesla and Waymo have strong proprietary self-driving systems, Nvidia is “doing it for everyone else.”
Lucid is planning to offer the first consumer-owned Level 4 autonomous vehicles by integrating Nvidia Drive AGX Thor into its upcoming mid-size platform.
Production of this model is scheduled to begin later this year in Saudi Arabia.
Besides its integration within Lucid and Mercedes-Benz vehicles — the Alpamayo solution was first tested with the German automaker’s models scheduled for release later this year — Huang announced on Monday that several Asian automakers will be building vehicles on Nvidia‘s Drive Hyperion program.
These include Japanese manufacturers Nissan and Isuzu, China’s largest auto companies BYD and Geely, and South-Korean Hyundai Group.
Competition to Tesla
According to Tesla‘s CEO Elon Musk, however, the plans are similar to what Tesla is currently doing with its Robotaxi service, which is powered by the Full-Self Driving (FSD) software.
He wrote on X last January that Nvidia‘s software does not pose “competitive pressure” on the company.
“The actual time from when FSD sort of works to where it is much safer than a human is several years,” he stated, adding that “the legacy car companies won’t design the cameras and AI computers into their cars at scale until several years after that.”
According to Musk, “this is maybe a competitive pressure on Tesla in 5 or 6 years, but probably longer.”
Musk reaffirmed late last year that Tesla was open to licensing its FSD software to other automakers; however, despite having tried to offer them the system, “they don’t want it.”
He had first suggested in 2021 that Tesla might license its FSD software to other automakers, noting that he had discussed the idea with several companies.
These discussions did not result in any agreements.
Meanwhile, Tesla continues to struggle with regulatory approval for its Full Self-Driving (supervised) system in Europe, despite repeated assurances from the company and its CEO over the past months that approval would arrive this quarter.
Robotaxi Landscape
Tesla is pushing forward with its intent to expand the Robotaxi service — launched late last June in Austin — across the United States.
On January’s earnings call, Musk said the company achieved “100% unsupervised” self-driving, as it removed the operators from its Robotaxi fleet vehicles and chase vehicles from its operations.
The company is scheduled to begin manufacturing its Cybercab model in April, after the first units rolled off the production line last month.
The purpose-built autonomous vehicle, unveiled in late 2024, will be used in the autonomous ride-hailing service.
Tesla uses a vision-only approach to self-driving, unlike major competitors such as Google-backed Waymo, which considers LiDAR integration to be the “safest path.”
Nvidia‘s Drive AV software fuses data from LiDAR, cameras, radar, and other sensors to give autonomous vehicles a comprehensive understanding of their surroundings.
Commenting on Waymo and Tesla earlier this week, co-founder and ex-CEO of Uber Travis Kalanick said that “Waymo is obviously ahead, despite having issues on “manufacturing, scale, urgency, and fierceness.”
On the other hand, Tesla has the “fundamentals, science, hard mode times 100.”
Despite acknowledging that the Elon Musk-led company’s approach is “super inspiring,” he noted that the question is how will the company “get there and in what timescale.”
“Honestly, everybody’s like, ‘could happen tomorrow, could happen in five years.’ And I think that, like, when does the ChatGPT moment happen for vision? It’s super inspiring, but what’s the timeline on it?,” he added.
During Monday’s event, Nvidia‘s CEO reiterated the idea that “the ChatGPT moment” for robotics and autonomous vehicles is here, something he had been stating since January.
