When self-driving startup Nuro emerged from stealth mode in early 2018, its ex-Googler cofounders had a simple reason for why they wanted to create street-legal delivery robots instead of autonomous cars or robotaxis: it’s a little easier and you can create a moneymaking business faster. As the coronavirus crisis rolls on, that looks like a particularly shrewd decision.
“We certainly didn’t hope that the pandemic would happen. It’s tragic,” Jiajun “JZ” Zhu, Nuro’s cofounder and CEO, tells Forbes. At the same time, it triggered a “3X” jump in demand for the company’s automated delivery services from partners like Krogers KRand food banks, and is preparing to work with Walmart WMT and Domino’s pizza. “Definitely we see that delivery, especially contactless delivery, is something that people really need, and we hope that we can do more here.”
Having raised $1 billion, much from SoftBank’s Vision Fund, the Mountain View, California-based company is positioning itself to do that. Initially overshadowed by bigger companies, closely-held Nuro has become one of the more intriguing players in the race to create viable businesses developing software to drive vehicles as well or better than humans and is a standout in the 2020 edition of Forbes’ AI 50: America’s Most Promising Artificial-Intelligence Companies.
Billions of dollars have flowed into dozens of startups and units at tech giants in the past four years to perfect the software, sensors and computing power needed for safe, reliable self-driving vehicles. Alphabet’s Waymo is the farthest along toward commercializing robotaxis, but technical challenges and health concerns triggered by the coronavirus have complicated the timetable for that business. As a result, Waymo is placing new emphasis on its Via automated logistics business as delivery and trucking services such as those being developed by Nuro and fellow AI 50 honorees TuSimple and, more recently, Aurora Innovation enter the fast lane. (Likewise, Amazon’s interest in robotaxi startup Zoox, which it’s planning to buy for about $1.2 billion, may be about leveraging its self-driving software for package delivery operations.)
“Moving stuff seems like it has many advantages over moving people in the short term for autonomy,” said Mike Ramsey, Gartner IT’s Detroit-based auto and smart mobility analyst. “It’s much easier to construct the business case and you can also more easily limit the domains where they operate.”
TuSimple, which has the biggest fleet of self-driving semi-trucks in the U.S., has been able to maintain revenue-generating deliveries throughout the pandemic. Even so, its trucks still have a human safety driver at the wheel and a technician monitoring system performance. By contrast, electric Nuro R2 delivery bots, half the size of a standard passenger car, carry only goods and are designed for urban and suburban streets, with a top speed of 25 miles an hour. Like passenger-oriented autonomous vehicles, the delivery bots are equipped with laser lidar sensors, radar and digital cameras to monitor surroundings. But occupant safety isn’t a concern, just the safety of others on the road–and keeping the eggs from breaking on grocery runs.
Nuro’s robots also don’t feel the need to speed human drivers contend with. “What that means in practice is if we’re entering an area where we see there are a lot of vulnerable road users, like pedestrians or kids or cyclists, we can be a little bit more conservative, drive a little bit slower, give them all a little bit more berth, in ways that your Uber UBER driver probably wouldn’t but that would overall increase the safety,” says cofounder and President Dave Ferguson. “Maybe that increases our trip time by 30 seconds, but the groceries don’t mind so much.”
The company won approval from the U.S. Department of Transportation in February to deploy R2s, which lack conventional driving controls, pedals, windshields and mirrors, and in April California okayed the toaster-like vehicles for use in a handful of Silicon Valley communities.
Modified Toyota Priuses, part of Nuro’s initial test fleet, are used for grocery deliveries mainly in Houston for Krogers as the R2 fleet is built up. Nuro won’t say exactly how many it currently operates. R2s were made available to haul medical supplies for temporary Covid-19 treatment centers in Sacramento and San Mateo, California in April. In May, Nuro launched a pilot in suburban Houston with CVS to deliver prescriptions and other items ordered online. Nuro generates revenue from these programs, but won’t say how much. By year-end, the company plans to have between 50 and 100 R2s operating in the U.S.
“When we look at the status of the technology, including both the hardware and the software, we are moving very quickly and are continuing to try to get as large a fleet out there doing real essential services as we can,” Ferguson says. Eventually, it could scale to hundreds and then thousands of R2s.
Neither Zhu nor Ferguson shared long-term revenue targets, but the company’s current valuation of $2.7 billion, according to PitchBook, suggests investors see significant potential.
Still, programs like Waymo’s robotaxis in Phoenix and large-scale projects by General Motors GM-backed Cruise in San Francisco and Ford- and Volkswagen-affiliated Argo AI in Pittsburgh and Miami, benefit from a big head start. “I expect autonomy applications in trucking and logistics will mature more quickly–even if passenger autonomy comes to market sooner,” Ramsey, Gartner’s auto and smart mobility analyst, said.