Waabi recently released a blog post about its latest business updates and how its autonomous trucks are literally “taking it to the streets.” It’s less an advocacy pun than a competitive advantage, as many autonomous big rigs have focused on highways. The blog post also included the announcement of a new direct-to-customer business model.
FreightWaves spoke with Lior Ron — founder of Uber Freight, its former CEO and now chief operating officer at Waabi — about the company’s recent developments and to catch up since he took over the role last August.
Ron noted that since taking the role, one thing that surprised him was how fast autonomy and software are advancing, in part because of rapid progress in generative AI. Autonomous trucking has gone from “just wait, it’ll be ready in two more years” to states crafting specific regulations for virtual drivers.
“It just feels mature, like we’re now in the next inning in terms of specifics, in terms of making concrete plans, in terms of buying, in terms of getting ready for deployment, in terms of being thoughtful and deep on how we’re thinking about the opportunity. It’s just so much more mature, thoughtful and progressed than what I expected,” Ron said.
These rapid advancements have also narrowed the gap to full commercial deployment. What took years now takes quarters as timelines compress. For Waabi, the focus on full commercialization goes hand in hand with Volvo, its OEM partner.
“We will be ready with our software autonomy solution soon, but we believe the only way to safely remove the driver is to have the truck and the fully redundant platform ready from the OEM and blessed by the OEM as well,” Ron said.
Ron added that the era of fully driverless operation will come once both sides of the hardware and software equation are ready. The hardware side — and when it will be blessed by the OEMs and fully ready — remains a key point of friction between virtual-driver software makers and the companies that build the vehicles those drivers will control.
Another point to resolve is the debate between point-to-point and hub-and-spoke models.
Point-to-point or hub-and-spoke?
In trucking, most shipments are point-to-point: A driver picks up at a shipper and goes directly to a receiver. This is simpler than hub-and-spoke, which less-than-truckload or parcel carriers use more often. In hub-and-spoke, one driver shuttles a trailer to a yard or terminal, where another runs the middle-mile linehaul route to another terminal. A final driver then takes the trailer to its destination.
To understand the future of autonomous trucks, one must understand how they will be used — and, more importantly, what customers are willing to pay for.
Ron and Waabi are firmly in the point-to-point camp, which they believe will unlock far more value than hub-and-spoke approaches from competitors. This view was shaped in part by feedback from potential customers.
“Highway is an okay start, but it has inherent issues that don’t make it a fit for real commercial deployment. I would say first and foremost it adds an additional cost on both sides that, as we both know, nobody’s willing to pay for,” Ron said.
Ron continued, noting that the added costs of highway-only plus local moves will be a pain point when negotiating rates and determining who gets paid. “So the added cost on both sides if you’re only doing highway … At the end of the day, somebody needs to pick up the freight from the origin and somebody needs to drop it at the destination. Those are additional costs of $200 to $300 on both sides — first- and last-mile — that nobody’s going to pay. By nobody, I mean the customer is not going to pay a single cent more.”
This can lead to what Ron calls “the highway trap,” a big problem because it raises questions about who subsidizes whom. For example: What rates should the carrier get for shuttling the trailer just to enable the middle-mile autonomous leg? Does this get charged back to the customer, or would the autonomous vehicle company bake an all-in rate that includes human drivers for the first- and last-mile legs?
Another issue Ron noted is service reliability when managing customer expectations in a hub-and-spoke system. It’s not like less-than-truckload, where the hubs and spokes are mostly under one carrier. An autonomous hub-and-spoke setup would require at least two other carriers to handle the first and last mile.
Ron gave the example of a consumer packaged goods customer with a shipment being delivered to a Walmart. For those less familiar with trucking operations, some customers like Walmart demand delivery within a specific window. Arrive too early and you may get a fine. Arrive one minute too late, and you get another fine.
Add the complexities of managing an on-time pickup, on-time terminal transfer and scheduled yard-to-receiver delivery, and you have three moving parts. Each increases the risk of failure in this three-legged shipment. It’s why many trucking companies prefer point-to-point, cutting out the middleman — and it’s hard enough to manage pickup and delivery on time even in perfect conditions.
“For those reasons, autonomy on highways only doesn’t really scale. It has room in the ecosystem — 5% to 10% or 15% of the future network will be hub-to-hub. But the vast majority of freight will be moved door-to-door, facility-to-facility, end-to-end. That’s how the supply chain is built today. This is how the supply chain is going to be built in the future as well. So I think for us this is a major unlock. I would say it’s very clear — the unspoken truth in the industry,” Ron said.
To beat hub-and-spoke, the autonomous truck must do it all, including navigating small side streets and local roads. It’s a much more difficult task, but the generative AI revolution has made it easier — or at least faster. Waabi believes it has an advantage here.
“Getting away from just highways is the next level in terms of technical difficulty. And that’s one of the benefits of the AI-first approach and what we’re doing with simulation and the ability of the system to generalize: We can actually escape the highway-only trap and, for the first time in the industry, really bring true end-to-end abilities and allow autonomy to happen across, at this point, the vast majority of road types and facility types to really unlock what commercial deployment of the future is going to be,” Ron said.
Ron also cited customer feedback pointing to door-to-door service. A Fortune 500 shipper doesn’t care about the autonomous challenges of smaller streets. To the customer, it’s about cost, service and scale. The carrier is a cog in the inbound and outbound transportation machine. “Autonomy providers need to meet customers where their supply chain, logistics and operations are. That’s the only way this is going to work,” Ron said.
Timelines and partnerships pave the way to deployments
Waabi’s initial deployment model will be with Volvo and its unit, Volvo Autonomous Solutions. The Swedish-based global OEM has taken a pragmatic approach to autonomous truck hardware development, deciding to operate as a trucking company and learn how to run it before mass-producing the vehicles.
Volvo Autonomous Solutions, or VAS, has the terminal, tractors, shippers and everything except the virtual driver, which is housed in a server rack slightly larger than a college dorm refrigerator. That in itself is a testament to how fast hardware has advanced. In the past, the robot brain of an autonomous truck took up the entire sleeper section of the cab — imagine the equivalent of three countertop-sized refrigerators in the bunk.
More processing power in smaller packages, plus sensors like radar, lidar and cameras that now contain tiny microprocessors directly in the device — a concept called edge computing — are key ingredients that have shortened autonomous trucking hardware timelines from years to quarters.
For Waabi, the partnership with VAS is a boon. “What’s nice about that — think about them as the first fleet to adopt autonomy. We can, together with Volvo, develop the customer model of how to deploy, how to participate in the routing guide, how to do all the operations autonomously on that route, how to do maintenance, how to ensure uptime, what the ROI is for those customers. We can do all of that,” Ron said.
“I would say that will be the early initial deployment model: basically using Volvo as a fleet with the Waabi driver to unlock autonomous lanes and deliveries on your network as a customer, whether you’re a 4PL, 3PL, shipper or whatnot. That’s how it starts,” he added.
After that, expect adoption rates to accelerate as VAS tests the maintenance, uptime and performance of running a driverless truck with enough redundant systems to make an aerospace engineer blush. That’s another reason OEMs are in less of a hurry: The redundant systems need testing, and there aren’t many real-world examples of a Class 8 truck with double braking, steering, compute, sensors and other subsystems. Current Class 8 trucks have a human driver, so there’s no need for such complexity.
Looking ahead, based on conversations with Waabi and other autonomous truck technology providers, 2026 will be the year of more firsts but also more refinement.
“So it’s our job — and by us I mean the industry, Waabi, Volvo, etc. — to put all of the pieces in place and reduce any unknowns or questions. We are doing that step by step: technology, truck availability and geography,” Ron said. “Facility-to-facility was a huge unlock because that was really stopping people from having deeper conversations — because if I can’t do it facility-to-facility, don’t force me to get it to the highway.”
While the technology continues to mature, the second development to watch will be hardware refinement by the OEMs.
“You want the OEM — they’re producing the truck. You also want the OEM, if you’re a fleet or a customer, to say, ‘Hey, we think this truck is ready,’ versus me as a software provider coming to the customer saying, ‘Hey, I vouched for the truck.’ Well, are you a hardware producer? No. Did you develop the truck? No. So how do you vouch for the truck? I think that’s a very important question that we see a variety of answers to today in the industry, and we definitely err on the side of caution and safety. We only want to remove a driver once the truck is fully ready,” Ron said.
