• Uber autonomous vehicle production to reach 100,000 units by 2027.
• To use Nvidia’s DRIVE AGX Hyperion 10 platform in collaboration with Stellantis, Lucid, and Mercedes-Benz.
• Combines Nvidia AI infrastructure with Uber’s marketplace to deploy of Level 4 robotaxis.

Uber and Nvidia have formalised an expanded partnership that could represent a turning point in autonomous vehicle commercialisation. The ride-hailing company announced this week it will deploy Uber autonomous vehicles at unprecedented scale – targeting 100,000 units starting in 2027 – using Nvidia’s DRIVE AGX Hyperion 10 platform as the technical foundation.

Unlike previous autonomous vehicle initiatives that relied on vertically integrated, proprietary systems, this approach establishes an open ecosystem framework designed to accelerate industry-wide adoption of Level 4 robotaxis.

The October 28 announcement signals a shift from isolated pilot programmes to platform standardisation, bringing together multiple automakers and technology providers under a common architecture. It’s a bet that modularity and interoperability, rather than closed systems, will unlock the scale needed to make robotaxis economically viable.

The technical foundation

At the heart of the expansion lies Nvidia’s DRIVE AGX Hyperion 10. It features dual Nvidia DRIVE AGX Thor system-on-chips built on Blackwell architecture, delivering a combined 4,000 FP4 teraflops of compute.

The sensor suite comprises of 14 high-definition cameras, nine radars, lidar, and 12 ultrasonic sensors, integrated with the Nvidia DriveOS and Nvidia DRIVE AV software purpose-built for Level 4 autonomy.

“Nvidia is the backbone of the AI era, and is now fully harnessing that innovation to free L4 autonomy at enormous scale, while making it easier for Nvidia-empowered AVs to be deployed on Uber,” said Dara Khosrowshahi, CEO of Uber.

Modularity distinguishes the new approach. Unlike proprietary systems that locked automakers into specific configurations, DRIVE AGX Hyperion 10 allows manufacturers to customise the platform and maintain compatibility in the ecosystem. Standardisation could prove crucial in accelerating industry-wide adoption.

A growing ecosystem

The partnership extends beyond a bilateral agreement. Stellantis will be among the first original equipment manufacturers to deliver at least 5,000 Nvidia-DRIVE-powered Level 4 vehicles to Uber for robotaxi operations in the United States and internationally.

The Italian-French automaker is developing AV-Ready Platforms specifically optimised for robotaxi requirements, collaborating with Foxconn on hardware and systems integration. Lucid is advancing Level 4 autonomous capabilities for its next-generation passenger vehicles using the full-stack Nvidia AV software on the DRIVE Hyperion platform for upcoming US models.

Mercedes-Benz is testing future collaboration powered by its proprietary MB.OS operating system and DRIVE AGX Hyperion, with the new S-Class offering what the company describes as an “exceptional chauffeured level 4 experience.”

Beyond passenger vehicles, the ecosystem will extend to freight. Aurora, Volvo Autonomous Solutions, and Waabi are developing Level 4 autonomous trucks powered by NVIDIA DRIVE AGX Thor.

Nvidia and Uber will continue to support the broader global Level 4 ecosystem, including Aurora, Avride, May Mobility, Momenta, Motional, Nuro, Pony.ai, Waabi, Wayve, and WeRide in passenger mobility, trucking, and delivery applications.

The data challenge

Perhaps the most technically significant aspect of the partnership is the joint robotaxi data factory, powered by Nvidia’s Cosmos platform for physical AI. Uber will collect more than three million hours of robotaxi-specific driving data to fuel Level 4 model training and validation.

“Together, these capabilities form a powerful data engine – spanning ingestion, labelling, scenario mining, synthetic data generation, and large-scale training – that aims to shorten the path from pilot to profitable autonomy deployment,” according to Uber’s press release.

Nvidia is also releasing what it claims is the world’s largest multimodal autonomous vehicle dataset, comprising 1,700 hours of real-world camera, radar, and lidar data from 25 countries.

Nvidia uses vision language action (VLA) models that combine visual understanding, natural language reasoning, and action generation. “By running reasoning VLA models in the vehicle, the AV can interpret nuanced and unpredictable real-world conditions – like sudden changes in traffic flow, unstructured intersections and unpredictable human behaviour – in real time,” Nvidia states in its release.

Safety and certification

Nvidia has launched the Halos Certified Program, which it describes as the industry’s first system to evaluate and certify physical AI safety for autonomous vehicles and robotics. The Nvidia Halos AI Systems Inspection Lab, dedicated to AI safety and cybersecurity, has received accreditation from the ANSI Accreditation Board.

Companies including AUMOVIO, Bosch, Nuro, and Wayve are among the inaugural members of the inspection lab. AV toolchain leader Foretellix is collaborating with Nvidia to integrate its Foretify Physical AI toolchain with Nvidia DRIVE for testing and validating these models.

Industry implications

The announcement represents a notable move toward platform standardisation in autonomous driving. Rather than each automaker developing proprietary systems from scratch, the Nvidia-Uber framework offers a common foundation that could accelerate time-to-market while potentially reducing development costs.

“Robotaxis mark the beginning of a global transformation in mobility – making transportation safer, cleaner, and more efficient,” said Jensen Huang, founder and CEO of Nvidia. “Together with Uber, we’re creating a framework for the entire industry to deploy autonomous fleets at scale, powered by Nvidia AI infrastructure.”

However, significant challenges remain. Regulatory frameworks for Level 4 autonomous vehicles vary dramatically across jurisdictions. Questions about liability, insurance, and operational safety standards are still being resolved. The path from 5,000 vehicles to 100,000 will require regulatory approval in multiple markets.

It’s worth noting that the economics of robotaxi operations remain unproven at scale. While eliminating human drivers reduces operational costs, the capital expenditure for Level 4-capable vehicles, ongoing maintenance, remote assistance infrastructure, and the data processing requirements for continuous improvement represent substantial investments.

The target of 100,000 Uber autonomous vehicles by an unspecified timeline starting in 2027 is ambitious. Whether this partnership can deliver on that promise – and whether the broader industry adopts this platform approach – will likely define the next phase of autonomous vehicle commercialisation.

What’s clear is that the autonomous vehicle industry is moving from isolated pilot programmes toward ecosystem-based deployment models. The Nvidia-Uber partnership represents one significant approach to scaling this technology, though its ultimate success will depend on execution, regulatory developments, and market acceptance in the years ahead.

Want to experience the full spectrum of enterprise technology innovation? Join TechEx in Amsterdam, California, and London. Covering AI, Big Data, Cyber Security, IoT, Digital Transformation, Intelligent Automation, Edge Computing, and Data Centres, TechEx brings together global leaders to share real-world use cases and in-depth insights. Click here for more information.

TechHQ is powered by TechForge Media. Explore other upcoming enterprise technology events and webinars here.