Autonomous Trucking Is Here, but Driver as a Service May Not Be

By Brett Suma

For most of the last decade, autonomous trucking has seemed like it’s just “two years away.” I don't believe that’s the case anymore. Autonomy is here. Trucks are already moving freight without drivers, the technology is rapidly advancing, and autonomous operating networks will continue to expand. However, I believe meaningful adoption of Driver as a Service is still many years away.

That view is informed by my regular conversations with executives at some of the largest trucking companies in North America. These are people I have known and worked with for years, and their message is remarkably consistent: they are watching autonomy closely, but they have little interest in being first movers. They want to see the technology operate safely, reliably and economically at scale before making significant commitments. When the conversation moves from the autonomous technology itself to how it would actually be deployed inside their businesses, their questions become much tougher.

For example, large carriers are unlikely to purchase autonomous tractors at meaningful scale unless they come directly from the OEMs they already rely on – complete with the warranties, parts, service networks and maintenance support they expect for the rest of their fleet. That means an ADS developer being technically ready isn't enough for a large carrier to fully invest in them. The OEM still has to manufacture and warrant the autonomous-ready tractor, the dealer and maintenance ecosystem has to support it, and the carrier has to be comfortable owning it.

And by most indications, the OEM ecosystem required to support this model at scale is still at least a couple of years away. Technology readiness and fleet readiness are two very different things, and a successful DaaS strategy requires both to work.

There is some irony that after spending a decade hearing autonomy itself was “two years away,” the technology has finally arrived—and now one of the critical dependencies for DaaS is, once again, two years away. Technology readiness and fleet readiness are two very different things, and a successful DaaS strategy requires both to work.

Then there is the economics. 

DaaS is often presented as an all-in autonomous-driver cost that can be compared with the cost of a human driver, but that isn't the full cost of operating an autonomous tractor. Removing the driver doesn't eliminate everything the driver does; much of it needs to be replaced with infrastructure that still carries a cost. Inspections, fueling, sensor cleaning and calibration, remote assistance, specialized maintenance, connectivity, roadside recovery, secure staging and physical exception management have to be done by someone. If the ADS provider doesn’t perform those tasks, the carrier will absorb those costs. Calling the autonomous driver “all-in” doesn't make the autonomous transportation system backing it all-in.

The same problem exists with the idea that DaaS is “asset-light.” It may be asset-light for the ADS developer, but for the carrier it means purchasing a more expensive autonomous tractor, supporting the infrastructure around it and assuming the risk of keeping it utilized. Yet unlike a conventional tractor, the carrier does not control where that asset can operate. Its utilization is constrained by an operating domain defined by autonomous technology rather than by the carrier's freight network.

That is a significant change in trucking economics. A traditional carrier can redirect a traditional tractor when freight changes, but an autonomous tractor can only go where the technology can operate. 

If the carrier's freight doesn't match that network, the asset can't simply follow the freight. The result could be a very expensive tractor, supported by equally expensive infrastructure, generating far fewer productive miles than the carrier originally modeled. The ADS developer can still describe its model as asset-light while the carrier is left with the asset and the utilization risk.

Still, there is an even bigger constraint at play: the carrier doesn't ultimately control whether autonomy gets used. The shipper does. A carrier can decide it wants autonomous tractors, but it cannot unilaterally decide that a customer's freight will move autonomously. The shipper controls the freight, the facilities and, ultimately, whether autonomous equipment is acceptable within its supply chain. 

From my conversations with shippers, I found that many of them are even more conservative than the carriers. They want to understand safety, liability, insurance, cargo security, facility operations, and contingency procedures before allowing autonomous equipment into their networks, and that’s the kind of language they are writing into their carrier agreements.

That creates a difficult position for a carrier considering DaaS. The carrier is being asked to purchase a premium asset whose operating domain it doesn't control, build or support infrastructure specific to that asset, and then depend on its customers to provide enough eligible freight to keep it utilized. 

At the same time, selectively removing freight from human-driven lanes can disrupt the carrier's broader driver network. A carrier doesn't operate one lane in a vacuum; it operates an interconnected system of drivers, tractors, trailers and customers. What looks economically attractive on one autonomous corridor can cause problems elsewhere in the network.

This is why Bot Auto has taken a different approach, offering Transportation as a Service through a compounding freight network. We don't believe trucking companies should have to completely transform in order to benefit from autonomy. 

The Capacity Grid Powered by Bot is designed to bring together the technology, tractors, infrastructure, freight and operations required to create autonomous capacity. Rather than handing a carrier an autonomous driver and asking it to solve everything around it, we take responsibility for building the system that makes the autonomous driver useful.

Just as importantly, we work directly with the companies that ultimately control the freight. So, instead of asking a carrier to buy an autonomous tractor and hope enough of its customers have freight that fits an ADS-defined operating domain, we aggregate demand from multiple shippers, carriers and logistics providers across the Grid. The infrastructure isn't being built around a single carrier's autonomous fleet; it supports a network where increasing numbers of customers, lanes and freight movements can share the same autonomous capacity. That allows us to attack the problem that matters most in the capital-intensive transportation business: utilization.

It starts with a different premise: don't sell the trucking industry an autonomous driver and ask it to figure out how to use it. Build the network that turns autonomous technology into capacity and sell the capacity.

I don't think the trucking industry is waiting another two years for autonomy. Autonomy is already arriving. What may take considerably longer is aligning everything DaaS requires around it: an OEM-manufactured and warranted autonomous tractor, a carrier willing to purchase that more expensive asset, infrastructure capable of supporting it, enough shipper-approved freight within an ADS-defined operating domain, and sufficient utilization to make the entire investment economically compelling.

That could take five years. It could take ten.

The Capacity Grid doesn't require us to wait for that transition.