Nuro Narrows Focus, Sets Sights on AV Application in 2026

Autonomous-vehicle developer Nuro has torn up its playbook.

The California-based developer, which launched eight years ago with a goal of producing driverless package-delivery vehicles, has changed course, scrapping plans to manufacture purpose-built vehicles and operating autonomous commercial fleets to focus strictly on licensing its proprietary fourth-generation Nuro Driver software to other vehicle manufacturers and service providers.

A casualty of the shift is a now-scrapped 2022 deal that had Nuro scheduled to mass produce its R3 (third-generation) package-delivery bots with China’s BYD, now the world’s leading seller of battery-electric vehicles.

If you live in Mountain View or Palo Alto, CA, or the Houston area, you may have seen some of Nuro’s low-speed bots delivering groceries for Kroger or takeout for Uber Eats.

Last month (Nov. 19), Nuro announced it would expand the reach of its test vehicles in those markets, increasing coverage 83% in the California territories and 70% in the Houston area, effectively opening up travel along all surface streets, including multilane roads. Operational speeds will be boosted up to 35 mph (56 km/h) to enable the vehicles to cover the additional ground. Nuro also will begin running its test fleets at night for the first time.

Although the operational advancements demonstrate the growing capability of the Nuro Driver software, the R3 vehicles no longer are part of the new game plan revealed in September.

“Earlier this year, we announced a change in our business model and effectively went from owning most of that value chain (for manufacturing and deploying driverless-vehicle fleets) to doubling down on our core competency, which is…trying to be the best AV tech provider that we can,” Nuro Chief Operating Officer Andrew Chapin tells WardsAuto.

In addition to selling its core software, Nuro will make the in-house digital tools it is using to develop Nuro Driver available for licensing by other AV developers as part of the company’s multi-pronged business approach that has it mining customers in the ride-hailing-service and last-mile package-delivery sectors, as well as automakers seeking L2-plus and L3 advanced driver-assistance system technology.

Prompting the switch in strategy was a realization Nuro couldn’t support its original broad remit.

“The general macro situation has changed quite a bit,” Chapin says. “The cost and availability of capital has necessitated companies like Nuro get very focused on what they do best, as opposed to taking on as much of the value chain as possible. (And) there are other players that are willing to take on some of what we previously thought that we had to own ourselves.”

Few companies have successfully operated driverless fleets in North America, he notes, and Nuro says it has covered more than 1 million real-world miles (1.6 million km) autonomously without causing a single accident. Although the vehicles are monitored remotely, Chapin says human control over the vehicles has occurred less than 1% of the time in the developer’s California and Texas testing.

Nuro’s perception stack relies on the typical trio of camera, lidar and radar sensors and claims to be among the first to deploy the powerful NVIDIA DRIVE Thor processor in an L4 driving application.

Chapin says the performance of the Nuro Driver software stack has advanced greatly in recent months due to huge leaps made in artificial-intelligence capability.

“We have really invested a lot over the last 12 to 18 months to have every part of our AV stack powered by machine learning and AI – so that's everything from mapping to perception to behavior,” the Nuro COO says. “And the rate of progress has really been tremendous over that period.

“(We’re getting) to the point where (Nuro Driver is) scalable and can operate in a broad enough environment that we can start to commercialize it soon.”

Penetrating the market won’t be easy. Many equally optimistic competitors are facing Nuro in the still-nascent AV-development space, including Waymo (the most successful so far in operating small driverless fleets in North America), General Motors’ Cruise, Tesla and a multitude of independent software developers, including NVIDIA, Mobileye, Baidu, Pony.ai, Zoox and others.

In the Level 2-plus to Level 3 ADAS market, Nuro will have to get through the door of legacy and startup automakers to win business away from such institutional suppliers as Bosch, Continental, Aptiv, Valeo and ZF, all firmly in the mix.

Chapin believes Nuro may have an edge, however, from its proven operational track record and what he sees as its cost and scale advantages.

“There are a very small number of players that have deployed driverless (vehicles) at a city scale,” he says. “We believe, with high conviction, that the overall technical maturity and sophistication of the Nuro Driver is right up there with anyone else in North America.”

The company’s new, sharper focus will allow it to move faster when it comes to deploying its software in production vehicles, Chapin adds.

 “If we were having to produce our own vehicles (and) operate the fleet, that becomes a headwind to scalability,” he says. “Licensing our tech into this ecosystem helps with making the model more scalable, because we can rely on other players whose core competency is vehicle production (and) vehicle-asset ownership and operation.”

As for cost, Chapin says Nuro is using less-expensive off-the-shelf sensors, adding, “You cannot make a case that the unit economics for a ride-hailing business work when the self-driving system costs $100,000-$150,000 or more. Our system is going to be an order of magnitude less than that.”

Tests already are under way to prove Nuro Driver’s viability on highways, performance that will be needed to win ADAS applications in personal-use vehicles and, ultimately, to increase functionality for mobility-service providers.

First commercial applications should come by late 2026, Chapin predicts, adding there has been “significant interest” from potential customers. Overall, he sees a total addressable market in the trillions of dollars.

“I think in the next five to six years, you're going to see a lot more L3 offerings within personally owned vehicles,” he says. “On the L4 side, you have…one player in particular (Waymo) that has already started to try to scale, and we hope to be not far behind. I feel very confident that in two or three years we will…have our system integrated at the production line and starting to scale.”

The company, which lists more than a dozen investors including SoftBank, Google, Fidelity and TigerGlobal, reportedly has secured $2.13 billion over five rounds of funding.