Some day in the next few years, if you’re on the right stretch of highway in America’s Sunbelt, you are likely to have the disconcerting experience of pulling alongside a fully loaded semi truck, glancing at the cab, and seeing no one behind the wheel at all.
Unless you look closely, the truck you’re likely to see will look very much like a regular big rig. It will still have a steering wheel—twitching, as if moved by ghostly hands. It will also have those oversize rearview mirrors trucks have, only these will be even more exaggerated in scale, since they will double as mounts for sensors—including radar, lidar, and cameras—that help the truck see things even an experienced human driver might miss.
This truck won’t be as smart or adaptable as a human, but it will have superhuman senses, and won’t need to rest. What’s more, it won’t be susceptible to many of the pitfalls that have made autonomy in passenger vehicles largely a disappointment, with companies blowing past one self-imposed deadline after another. While the self-driving passenger-vehicle industry struggles to gain traction despite decades and tens of billions of dollars in investment, proponents of self-driving trucks say they could be here—and making money for their operators in commercial services—much sooner.
Some of the companies involved say they will have the first trucks without drivers in the cab on America’s highways by the end of next year. Those include
which has partnerships with
which has joined up with
When it gains widespread traction, robot trucking will have big implications for how we move goods around America—and for the companies and people involved in that process. For starters, it could help alleviate a chronic shortage of drivers, who are retiring faster than they can be replaced, leading to what the American Trucking Associations claims is a historic shortage of 80,000 drivers.
Here’s the promise of robot trucks: While full self-driving in all conditions is still a pipe dream, engineers seem to be close to achieving it in limited circumstances, such as on highways on clear days. And highway driving, in good weather, happens to be exactly the context in which long-haul trucks operate for a substantial portion of the time.
One reason for that: Highways are what Aurora Chief Executive
“Another way to put that is that a bit of freeway in Texas looks very much like a bit of freeway in Phoenix or Minnesota,” says Mr. Urmson, a former faculty member at Carnegie Mellon University and Google executive who co-founded Aurora in 2017. The similarity is good for the artificial-intelligence technology that underpins self-driving, which can be great at handling things it has seen before, and terrible at adapting to situations that are novel. Anyone who has tried GM’s Super Cruise,
ProPilot Assist, or
Autopilot system has experienced this firsthand.
Highways also have the virtue of being relatively free of pedestrians, bicyclists, animals and children chasing after balls, and they tend to be well-marked and well-maintained.
Highways in Southwestern states, where the weather is generally good, are where autonomous trucking companies are currently testing their systems, carrying real loads for actual clients like FedEx and UPS, albeit with safety drivers behind the wheel in case the AI systems make a mistake—which they still do.
So, in good weather, a robot truck will see farther than a person can. It will never grow drowsy or inattentive. It will be able to operate 24 hours a day, stopping only for fuel and maintenance.
Now, there are also reasons to discount claims about commercial autonomous trucking happening in the next couple of years, or making a big difference anytime soon.
For one, the younger companies trying to pioneer the technology have to sustain the interest of investors until they start making money.
And that could be a while: Aurora, for example, has said that it will lose money until 2027. Shares in the company, which went public via SPAC in November 2021, are currently trading at about one-quarter of the $10 they sold for at their debut.
SHARE YOUR THOUGHTS
How comfortable do you feel with the prospect of sharing the road with autonomous big rigs? Join the conversation below.
Aurora Chief Financial Officer Richard Tame said the company has enough cash and short-term investments to fund operations through the introduction of its first autonomous truck next year and into 2024. The company has said in public filings that it expects it will eventually need to raise additional capital.
Even if startups do get their trucks rolling on schedule, it could take a while to have a real impact. By the end of 2023, Aurora will be putting “on the order of dozens” of driver-free trucks on America’s highways, says Sterling Anderson, the company’s chief product officer.
TuSimple also aims to have a fully autonomous commercial-trucking service operating in the U.S. by the end of 2023, says CEO Xiaodi Hou. In the meantime, the company also plans to begin delivering freight for Union Pacific with fully autonomous trucks, says a company spokesman.
Compared with the total number of large trucks rolling in America today—nearly four million, half of them the type that haul freight long distance—the scant dozens of self-driving trucks projected to be on the road by the end of 2023 would be a drop in the ocean.
Waymo—which, as a unit of Google parent
has less pressing concerns about funding—is less aggressive in its prediction for the arrival of robo-rigs. Its trucking-focused arm, Waymo Via, hasn’t set a date for its trucks to operate with no human in the cab, despite having already entered partnerships with trucking companies C.H. Robinson and J.B. Hunt, fleet-services operator Ryder, freight-brokerage company Uber Freight, and truck maker
Waymo has many reasons for that reticence, says Charlie Jatt, its head of commercialization for trucking. An important one is that there is no production-ready, commercially available truck with the redundant control systems that a self-driving system would require.
If power steering goes out in a human-controlled vehicle, a driver could still potentially muscle it to the side of the road. But with no human in the cab, an autonomous vehicle must have backup steering, braking and electrical systems, says Mr. Jatt. Getting all of these into trucks that can be made not just one at a time but by the tens of thousands is why Waymo has joined with Daimler Truck, he adds.
Indeed, everyone I interviewed for this piece, except for TuSimple, said that the potential for their systems to make mistakes is the reason they haven’t rolled them out yet—even those who claim to be close to doing so. (A recent report from the U.S. National Highway Traffic Safety Administration listed more than 100 accidents involving vehicles equipped with autonomous driving systems over the past year.)
Still, the potential financial benefits of robo-trucking technology are so enormous that shippers and trucking companies are likely to embrace it as soon as they feel it is ready.
“One thing that really surprised us was that the additional cost of the technology required by autonomous trucks is relatively small,” says Parth Vaishnav, an assistant professor at the University of Michigan and co-author of a recent study on the impacts self-driving trucks would have on truckers.
Adding even $20,000 of hardware, in the form of additional sensors and powerful computers, to a long-haul truck is quickly offset by the elimination of labor costs, which typically represent 15% to 20% of the cost of operating a truck. Another big economic impact is that by law a human driving a truck must stop and rest. That means every truck, which can cost between $100,000 and $200,000, is only being used about 30% to 40% of the time. Just running them 24 hours, stopping only for fuel and maintenance, increases their utilization by a factor of two or more.
“It’s economically so compelling that, even if other things about the truck modestly increase costs, it may turn out it will still be attractive,” adds Dr. Vaishnav.
Waymo’s strategy is to create a foundation of software and custom-built sensors—what the company calls the Waymo Driver. With some modification, that system can operate any vehicle autonomously, says Mr. Jatt, whether it’s giant semi-trucks or the autonomous taxis the company is currently operating in three different locations in the U.S. That might mean the company won’t be the first to deploy fully autonomous trucks, but the idea is that once it does start to roll them out, it has the potential to grow much faster than competitors, he adds.
Regulations are hardly a barrier to rolling out autonomous trucks. NHTSA has created suggested guidelines, which most states have adopted. These call for companies to, in essence, self-regulate. “A small handful of states have certain certifications or prohibitions, but all the rest are open for business,” says Aurora’s Mr. Anderson.
Autonomous-truck companies like his have for the most part moved on to validating their tech, rather than continuing to build it, says Don Burnette, CEO of Kodiak Robotics, another self-driving trucking startup. That’s why, for the industry as a whole, deployment of safe, commercially viable, fully autonomous trucks is just a couple of years away, he adds.
Longer term, robot trucks could go from driver-shortage solution to driver-job killer. There are approximately 500,000 truck drivers in America, says Dr. Vaishnav. Long-haul trucking jobs are among the most taxing and difficult, but also the best-paid jobs in the trucking industry.
In his model, deploying these trucks just in Sunbelt states could wipe out 10% of the total number of hours long-haul truckers in America spend on the road. That would mean the elimination of 30,000 to 40,000 jobs, says Aniruddh Mohan, a Ph.D. candidate who co-wrote the paper with Dr. Vaishnav.
As self-driving trucks become more capable, and can perform most of the driving on long trips throughout most of America, they could ultimately threaten nearly all long-haul trucking jobs, he adds.
Many in the automated-trucking industry claim that these job losses will be more than offset by the creation of new jobs, as self-driving trucks make trucking so much cheaper and faster that more freight shifts to it, at the expense of rail and even airfreight.
The trucks will all be remotely monitored by a human, and will need to be able to pull over on their own and signal for help. Beyond that, the specifics of human involvement are likely to vary widely from company to company, including whether a human driver will pilot vehicles in places like city streets.
It’s worth noting that existing long-haul trucking jobs are already a far cry from the solid middle-class gigs they were in the 1970s, before deregulation of the trucking industry, says Steve Viscelli, an expert in the trucking industry, sociologist at the University of Pennsylvania, and a member of the industry advisory council at Aurora. He believes self-driving trucks will continue what has been a decadeslong transformation of the industry.
“We may see autonomous trucks on the road in a couple of years, and I think after that the pace of adoption could be significantly faster than people expect, over the following decade,” says Dr. Viscelli. “These trucks are going to have different capabilities than a human-driven truck will have, so they will not be used in the way a human-driven truck will be used—in the same way a hundred guys with shovels are not an excavator.”
For more WSJ Technology analysis, reviews, advice and headlines, sign up for our weekly newsletter.
Write to Christopher Mims at email@example.com
Copyright ©2022 Dow Jones & Company, Inc. All Rights Reserved. 87990cbe856818d5eddac44c7b1cdeb8