Uber and Nuro have begun a new phase of testing for a premium robotaxi service in San Francisco, using Lucid Gravity SUVs outfitted with Nuro’s self-driving system. According to TechCrunch, select Uber employees can now request rides in these vehicles through the Uber app, with the vehicles operating in autonomous mode while a human safety operator remains behind the wheel as a backup.
This step is positioned as a progression toward a planned public launch “later this year,” but it is still limited to employee rides. For the robotics and autonomous vehicle industry, the test matters because it links together a full stack—vehicle sensing hardware, an autonomy computer, and rider-facing workflow—inside a live ride-hailing environment.
Employee rides as a live system test
TechCrunch reports that the San Francisco sightings may involve Uber employees riding in a Lucid Gravity SUV “blinged-out with sensors” and running a self-driving system developed by Nuro. The service is being tested via the Uber app, where select employees can request a ride in the robotaxi.
Nuro told TechCrunch that the vehicles are operating in autonomous mode while also keeping a human safety operator behind the wheel as backup. That combination reflects a common testing pattern: the autonomy stack drives the vehicle, but the presence of a safety operator provides a fallback during early operational validation.
In its own framing, Nuro said the employee test rides help evaluate how the autonomy stack, the vehicle, and the rider experience work together in a live operating environment. TechCrunch also notes that this setup allows Nuro to test how the vehicle handles rider pickups and drop-offs—an operation Nuro describes as “notoriously tricky” in autonomous ride-hailing.
From a technology perspective, pickups and drop-offs are an integration problem: the autonomy system must manage not only driving and perception, but also the interaction between a moving vehicle and a real-world rider workflow. Testing those moments inside an app-driven service can reveal failure modes that closed-course testing may not fully surface.
The hardware and compute stack: Lucid Gravity + Nvidia Drive AGX Thor
The robotaxi platform being tested is based on Lucid’s Gravity SUV, which was revealed in January, and is outfitted with sensing and compute designed to support real-world perception. TechCrunch reports that the Lucid Gravity robotaxi includes high-resolution cameras, solid-state lidar sensors, and radars. These sensors feed the self-driving system to perceive the real-world environment and operate in it.
Under the hood, the autonomy system is powered by Nvidia’s Drive AGX Thor computer, according to TechCrunch. The combination of multiple sensor modalities and a high-performance on-board compute platform is central to how autonomous systems handle dynamic environments: cameras and lidar can support detailed scene understanding, while radar can contribute to perception under varied conditions.
TechCrunch also links the robotaxi hardware choices to the operational goal of a ride-hailing service. Because the vehicles are being modified specifically for autonomous operation and sensing, the test is effectively a validation of the end-to-end system: sensors capture the environment, the Drive AGX Thor computer runs the autonomy stack, and the resulting driving behavior must work with the Uber app’s pickup and drop-off flow.
Funding, partnership momentum, and what’s scheduled next
While the current phase is employee-only, TechCrunch frames it within a broader timeline of partnership and investment. In July 2025, Uber announced a partnership with Nuro and a multimillion-dollar investment. TechCrunch reports that Uber invested $300 million in Lucid and separately agreed to buy “at least” 20,000 of Lucid’s new Gravity SUVs over the next six years.
That purchase agreement is relevant to the robotaxi testing because those vehicles are described as being equipped with Nuro’s autonomous vehicle system. TechCrunch also reports that Uber invested an undisclosed “multi-hundred-million dollar” amount into Nuro.
Looking forward, TechCrunch says production of these modified Lucid Gravity vehicles is expected to begin in late 2026, based on a regulatory filing posted last year. The planned public launch is described as occurring “later this year,” which suggests that the employee test is intended to bridge the gap between engineering readiness and a wider deployment.
TechCrunch further states that the plan is for Uber to own and operate the premium robotaxi service, likely with the help of a third party. That ownership and operations model could affect how the company scales: it implies Uber would be responsible for running the service, while hardware and autonomy may be developed and supplied by partners.
From closed-course testing to a multi-city engineering fleet
Nuro’s testing progression is also part of the story. TechCrunch reports that Nuro completed closed-course testing and started its first public road testing of the autonomous Lucid Gravity SUVs in late last year.
Nuro now has 100 Lucid Gravity SUVs outfitted with its self-driving system in an engineering fleet. TechCrunch says the fleet is used to gather real-world data and test autonomous driving across multiple U.S. cities and states. In this context, the San Francisco employee rides can be read as an additional layer of validation: it uses a live ride-hailing workflow rather than only road testing routes.
Because Nuro’s fleet is already operating across multiple locations, observers may watch for whether San Francisco introduces different operational constraints—such as pickup patterns, curbside behavior, and rider interaction—that require adjustments in how the autonomy system handles real-world logistics. The source does not provide specific results from San Francisco testing, but it does describe the purpose: evaluating the combined autonomy stack, vehicle behavior, and rider experience.
Why this integration test could matter for robotaxi deployment
Robotaxi programs often face more than just driving autonomy. TechCrunch’s details emphasize that the test is designed to evaluate how the autonomy system functions together with the rider experience, especially during pickups and drop-offs. If that integration performs reliably, it could support the path toward a broader launch; if problems appear, it could indicate where the autonomy stack or operational workflow needs refinement.
The source also ties the service to a specific compute and sensing configuration—high-resolution cameras, solid-state lidar, radars, and Nvidia Drive AGX Thor—suggesting that scaling the service may depend on replicating that hardware and compute setup across production vehicles. TechCrunch’s mention of expected production beginning in late 2026 provides a milestone that could influence how quickly the company can move from limited testing to larger deployments.
At the same time, the employee-only phase indicates that Uber and Nuro are still in a stage where human oversight remains part of the safety model, with a safety operator behind the wheel as backup. The source does not state when that backup requirement might change, but the current approach reflects how technology teams typically manage risk while expanding operational scope.
Source: TechCrunch