Mapping Uber’s robotaxi technology advancements

Uber initially pursued its own autonomous technology through its Advanced Technologies Group (ATG). ATG was established in 2015 by recruiting leading robotics experts and securing substantial investments, including a US$1 billion funding round from SoftBank, Toyota, and Denso in 2019. 

The goal was to build a proprietary autonomous vehicle system to replace human drivers, reduce costs and lead the market. However, after heavy financial losses, a fatal accident and increased scrutiny, Uber changed course. In December 2020, it sold ATG to autonomous vehicle startup Aurora and invested US$400 million in the company, gaining a 26% ownership stake.

Following the ATG phase, Uber transitioned from developing its own autonomous vehicles to establishing itself as a central platform connecting autonomous technology developers, automakers and specialized delivery companies. 

This model positions Uber as the distribution layer—the application for accessing transportation and robotaxi services regardless of vehicle brand. The company aims to serve as the primary operating system for autonomous mobility, leveraging its global scale, logistics expertise and extensive customer base to attract partners. Uber has secured agreements with more than 25 companies worldwide across the freight, delivery and robotaxi sectors.

Uber has formed partnerships with Chinese companies Baidu, Momenta, and Pony.ai; sidewalk delivery robot makers Cartken, Starship, and Serve; robotaxi developers AVride and Motional; as well as autonomous truck developers Torc and Waabi. 

Currently, Uber's autonomous ride-hailing services in the US are available in Atlanta, Austin and Phoenix, where Waymo robotaxis operate via the platform. Last year, Uber launched robotaxi rides in Dallas in partnership with Avride and in 2026 in Las Vegas with Motional, Hyundai’s autonomous driving arm. Internationally, Uber has teamed with WeRide to provide autonomous ride-hailing services in Abu Dhabi, Dubai and Riyadh.  

Through this model, Uber aims to become the foremost marketplace for autonomous transportation, transferring the substantial costs and risks associated with autonomous vehicle development to its partners while maintaining a crucial industry role in the future of mobility. 

By early 2026, Uber began leaning into a hybrid asset model, maintaining its role as a platform integrator while selectively investing in vehicles and infrastructure to secure supply. 

This shift was marked by two major deals: a US$300 million investment in Lucid Group for at least 20,000 autonomous Gravity SUVs with Nuro technology and a US$300 million deal between Uber and Rivian to obtain 10,000 R2 SUVs, with the possibility of investing up to US$1.25 billion by 2031 and expanding the order to 40,000 vehicles if certain autonomy targets are achieved. Uber also earmarked more than US$100 million for dedicated charging hubs intended to underpin future robotaxi fleets. 

Uber’s updated robotaxi strategy includes direct vehicle procurement, infrastructure investment and new operating costs. Scaling autonomous fleets requires significant upfront investment compared with traditional ride-hailing, which can grow by onboarding more drivers. 

By selectively investing in its own supply, Uber reduces the risk of vehicle shortages and maintains its central role in the autonomous vehicle ecosystem without embracing vertical integration. Uber is laying a foundation for offering a variety of vehicles and robotaxis tailored to user needs. 

The shift prioritizes a consistent, high-quality rider experience by influencing vehicle design through partnerships like the global robotaxi project with Lucid and Nuro.

Uber’s move to purchase and operate its own robotaxi fleet—despite the considerable upfront and ongoing investments—may yield significant long-term advantages. By owning autonomous vehicles, Uber will no longer need to allocate a large share of ride revenue to driver wages—a cost that has historically weighed on its profitability. 

The use of electric vehicles in the fleet further has the potential for inherently lower maintenance and energy costs versus a vehicle with a traditional internal combustion engine (ICE) or hybrid powertrain. Controlling infrastructure allows Uber to maximize vehicle usage, keeping its autonomous vehicles running nearly around the clock for both passenger and delivery services, unlike human drivers who may log off during slow periods. 

Uber’s decades of ride and user data provide an operational advantage that competitors would find difficult to replicate quickly and it may be used to more effectively manage a fleet’s uptime.  

No robotaxi is likely to be immune to deadhead time — the time between fares when the robotaxi has dropped a rider and is going to pick up the other — and there are hours in the day and night when ridership and transportation demand are low, but Uber’s data may enable the company to better navigate those periods. 

Despite high upfront costs, Uber’s operational expenses are expected to decline as its autonomous fleet matures, improving profitability despite vehicle ownership and maintenance costs.

Uber’s expansion has benefited from collaborations with tech companies and manufacturers worldwide as it targets launching robotaxi services in more than 10 markets across at least 15 cities—about half in the US and half internationally—by the end of 2026. Uber intends to grow its global autonomous fleet to 100,000 vehicles by 2027. 

Initiatives for 2026 include expanding autonomous services to Hong Kong, Houston, Madrid and Zurich. In the second half of 2026, Uber will pilot robotaxis in Tokyo with Nissan and Wayve. A commercial robotaxi service is set to debut in Zagreb, Croatia, with Pony.ai and Verne. Uber and Volkswagen are joining forces to introduce fully autonomous, electric ID. Buzz AD vehicles to the Uber platform in the US, beginning in Los Angeles in 2026.  

A cornerstone of Uber’s expansion of driverless robotaxis is its partnership with NVIDIA to build the world’s largest Level 4-ready autonomous mobility network using NVIDIA DRIVE AGX Hyperion 10, DriveOS, and the complete NVIDIA DRIVE software stack. Fleets powered by NVIDIA software are scheduled to roll out in 28 cities across Asia, Australia, Europe and North America by 2028, starting with Los Angeles and the San Francisco Bay Area in 2027. By 2029, Uber aims to become the world’s largest facilitator of autonomous vehicle trips. 

In addition to launching robotaxi services, Uber plans to leverage its large fleet operations as a valuable data resource by creating a robotaxi data factory using NVIDIA Cosmos. Over 3 million hours of specialized driving data will be collected to train and validate advanced autonomous vehicle models.

Uber has also established AV Labs—a dedicated unit focused on capturing complex, real-world driving scenarios and transforming operational data into training material for autonomous systems. Uber Autonomous Solutions supports partners in scaling and commercializing autonomous vehicles worldwide. It provides resources for infrastructure, user experience, and fleet operations to enhance reliability for passengers and efficiency for operators.