Waymo is the undisputed leader in the robotaxi field, operating a fleet of more than 3,000 driverless cars in at least 10 cities across the US. Several companies, including Tesla, Zoox, Everide and Motional, are in the race to compete with the Alphabet-owned company. But what if being No. 2 was actually better?
Nuro, a delivery robot company created by veterans of Google’s self-driving car project, believes it has a good chance to capture this niche. After moving from delivery to robotaxis in 2024, Nuro struck a deal with Uber and Lucid to deploy thousands of robotaxis across the US – receiving a multimillion-dollar investment from Uber itself in the process. Nuro plans to launch service in San Francisco later this year. And earlier this month, it was granted the first of many permits needed to launch that service.
Dave Ferguson, Nuro’s co-founder and co-CEO, said it’s almost better for Nuro that Waymo is moving at that pace. Its early successes, as well as its stumbles and mistakes, then became fodder for Neuro’s engineers to reexamine and reevaluate, with the goal of answering the question: Could we have done better?
What if being No. 2 was actually better?
“There is a lot of value in this kind of classic second mover perspective,” Ferguson said in a recent interview. “We have a lot of respect for Waymo … in some rare cases where they’re facing challenges, [Nuro is] “By using them we put pressure on our system and ensure that it will behave in a way that we are comfortable with and proud of.”
The fact that Ferguson has respect for Waymo is no surprise. He got his start in Google’s self-driving car project, which became Waymo, with Nuro co-founder and co-CEO Jiajun Zhu. The two left Google in 2016 to establish Nuro, first as a robot delivery service, and now as an aspiring robotaxi operator. Nuro aims to license its autonomous driving technology to outside companies, including car companies that want to use it for advanced driver-assistance systems and personally owned autonomous vehicles — though it has not announced any deals yet.
Sure, Neuro is late to the robotaxi party. While Neuro was handling groceries, Waymo was handling passengers. But Ferguson argues that Nuro’s technology could be easily transferred to robotaxis – even if its passenger experience is still nil.
This is where his “second mover” theory comes into play. Unlike Waymo, which had to figure out a number of operational challenges head-on, Nuro believes it will benefit from seeing the Alphabet-owned company operate a robotaxi service on a large scale before fully launching its own.
To that end, Ferguson wants Nuro’s robotaxi service to be widely useful when it first launches. He suggested that some features like freeway driving could come later, but the launch would not follow an ultra-incremental playbook where the company initially handles only certain scenarios before adding more complex scenarios over time. That said, Ferguson said, Nuro doesn’t plan to cover “the entire South Bay on day one.”
“The scheme has become a very useful service from day one,” he said. “It’s not just going to be protected intersections, and then gradually we’ll add unprotected… It’s going to be much broader [operational design domain] For starters.”
The Uber-Lucid-Neuro robotaxi service is unique because it involves three different companies: a rideshare network, an automaker, and an autonomous vehicle startup. Under the arrangement, Neuro develops the sensing and compute stack and works with Lucid to integrate that technology into the vehicle, the Lucid Gravity SUV. The integration happens directly on Lucid’s production line, meaning vehicles leave the factory already equipped with Level 4 autonomy. Those completed vehicles are then sold to Uber, which becomes the owner and operator of the fleet. The rideshare company will manage the depot and operational infrastructure associated with running the service.
Ferguson said Uber will also manage remote assistance for the vehicles. Remote assistance has come under criticism recently, with some members of Congress calling on Waymo and others to be more forthright about their use of offsite workers to maintain vehicles. Ferguson said this has led to misinformation about companies using remote workers to actively control robotaxis. What they really do, he said, is answer questions and provide prompts to help vehicles when they get confused.
“When people are told about remote assistance of self-driving vehicles they probably jump around like someone is driving a car in a dark room as if they were playing a video game,” he said. “I think this is a far cry from how remote assistance typically works.”
Nuro’s long-term goal, he said, is to build the most capable AI driving systems that are intended to be put to use in a variety of ways, including delivery. To that end, Nuro’s longevity in the field ensures that the company can apply lessons from its older, more rule-based machine learning systems as well as its current end-to-end learning models that produce a more natural driving style. According to Ferguson, this legacy remains important, even as the industry moves toward a more AI-heavy approach.
“You can think of it as a sanity check to make sure that what we’re doing is not too close to pedestrians, not too close to other vehicles, not violating any traffic rules,” he said.
He acknowledged that robotaxis suffer from a lack of public trust, especially in edge cases and other incidents where autonomous vehicles block traffic. Nuro intends to follow Waymo’s model of being transparent with some of its driving data in the interest of building trust with its customers, Ferguson said.
“The more evidence we have of Nuro and Uber and Lucid providing a product that is dramatically safer and better for our roads than a human-driven vehicle… the better it is for everyone,” he said. The company is still “trying to strike the right balance of how much detail we provide, so that it’s really understandable and relevant to the general public.” But Ferguson said he’s confident they’ll get there.
<a href