Autonomous heavy goods vehicles - an explainer

How will Britain benefit from autonomous HGVs? Why is their deployment behind robotaxis? And what is the state of autonomous lorries globally?

One of the most profound ways autonomous vehicles will affect day-to-day life in Britain is through the transport of goods, not people. Most domestic freight is carried by heavy goods vehicles, or HGVs1. In fact, about 80% of the total domestic freight moved in the UK is done on the road. So, how will Britain benefit from autonomous HGVs? Why is their deployment behind robotaxis? And what is the state of autonomous lorries globally?

Why autonomous heavy goods vehicles are valuable

Autonomous HGVs may be one of the most economically valuable uses of autonomous vehicles because they change how intensively the freight network can be used. A lorry is an expensive asset, but today its usefulness is constrained by driver hours, traffic and the inability to position the lorry in the most profitable haulage site when needed.

The first gain is fewer empty miles. In the UK, HGVs are empty for around 30% of the time they are driving. Autonomous HGVs would not remove empty running entirely, but they would reduce one important cause of it: the need to return a driver and vehicle to the same base, even when there is no profitable cargo to carry back.

Once that constraint is weakened, vehicles can be routed around freight demand rather than around a driver’s shift. A lorry could take a profitable outward journey, pick up a smaller return load, or follow a triangle between three locations where each has goods to send and receive. The industry already uses versions of this logic through trihaul routing, return-load platforms and groupage, where smaller shipments are combined into a single lorryload. This matters because groupage is already the largest category of UK HGV freight, accounting for 24% of goods moved in the UK, so autonomous HGVs would be building on an existing logistics pattern rather than requiring an entirely new one.

The second gain is time. UK lorries are active only for part of the day because drivers face legal limits on shifts and driving hours. An autonomous HGV could be used for much longer stretches, including overnight, when roads are quieter and journeys can be faster. A route that currently needs rest breaks could become a continuous journey, or a vehicle could deliver during the day and reposition itself overnight for the next high-value load.

This does not mean the system becomes frictionless. Goods still need to be loaded and unloaded, depots still need staff, and receiving sites still need to be open. But even with those bottlenecks, higher utilisation means each lorry can move more goods over its lifetime, with fewer wasted miles and less downtime.

The largest effect is that routes that are not profitable today become viable. Lower freight costs make it cheaper to serve rural areas, smaller towns and lower-volume routes. Faster and more flexible deliveries also matter most for goods where timing is valuable, such as food. If HGV freight becomes cheaper and more reliable, the gains do not stay inside logistics firms: businesses get lower transport costs, consumers get cheaper and more dependable goods, and places that are currently harder to serve get better access to supply chains.

This is also why autonomous HGVs provide an employment win. Thanks to induced demand, when it becomes cheaper and easier to move goods, people move more goods. The short-term adoption of autonomous HGVs is likely through a hybrid network, with autonomous HGVs handling predictable trunk routes while human drivers continue to serve complex local journeys. The new jobs will not all be driving jobs. They will appear in depots, warehouses, maintenance, remote operations and in the businesses that can open or expand once more places can receive goods faster and more cheaply. Most of the economic value would come from creating freight services, supply chains and local economic activity that do not currently exist.

Why are autonomous lorries behind robotaxis?

So if the economic benefits are so widespread, why are autonomous HGVs behind robotaxis in deployment? The answer is partly commercial and partly technical.

The economics favour robotaxis first

Waymo once had an autonomous freight division, and Google saw freight as a serious opportunity. But the programme was shut down in 2023. That was partly because robotaxis were already proving easier to commercialise and are more profitable per mile. Freight contracts may be larger, and the wider economic benefits of autonomous lorries may be greater, but robotaxis can earn more per mile in simpler operating environments: short urban trips, lower speeds and passengers willing to pay a high price for convenience. At the same time Waymos primary robotaxi competitor at the time, Cruise, was expanding quickly in San Francisco, putting pressure on Waymo to defend its lead in robotaxis rather than split attention with freight.

HGVs are harder to automate

But the gap is not only about business models. HGVs also create harder technical problems. These increase the risks of developing the technology and raise R&D costs. There are four main reasons: physics, the higher bar for reliability, the difficulty of training for edge cases, and the challenge of integrating an autonomous lorry into a system that still involves many humans.

Autonomous heavy goods vehicles have much longer stopping distances than robotaxis because they move faster and are heavier, so sensors need to detect risks from much farther away and react faster than a robotaxi going 20 miles per hour does.

That turns long-range perception into a central problem. Lidar works by sending out laser pulses and measuring what comes back, but distance makes the picture both thinner and weaker. To take an illustrative example: at 10 metres, two beams might land 3cm apart on the road, dense enough that a vehicle is hit by dozens of points and is unmistakably a car. At 150 metres, the same beams can land closer to 10 metres apart, wide enough that a smaller object becomes much harder to classify and, in some cases, easier to miss.

Range also weakens the signal. A reflective object, like a road sign, can still return a strong signal from far away, but a pedestrian in dark clothing, spray from a wet road or a dirty road surface may return much less. This matters because a heavy lorry travelling at 56mph needs over 100 metres just to stop in an emergency, and more distance still to brake smoothly. By the time hazards need to be detected 150-200 metres ahead, lidar’s point density has thinned and its return signal has weakened.

This is why autonomous lorry systems usually combine lidar with long-range radar and cameras. The challenge is not simply seeing the road. It is seeing far enough ahead, with enough confidence, for a heavy vehicle to react safely.

Beyond the physics problem, there is also a reliability problem. A robotaxi breakdown might mean a passenger loses half an hour. A failed autonomous lorry can delay an entire shipment, disrupt a depot schedule, and trigger thousands, sometimes tens of thousands of pounds, in penalties or lost revenue.

An edge case is a rare scenario that the system may not have seen often enough to handle confidently. This is hard enough in robotaxis: Waymo says its latest safety analysis covers more than 220 million fully autonomous miles, yet its cars have still needed recalls after encountering awkward real-world situations such as construction zones and gates. For HGVs, the long tail is even less forgiving. A burst tyre, a strange manoeuvre by another driver or a trailer behaving unexpectedly challenge the reliability of the service. Because the vehicle is heavier, slower to stop and harder to manoeuvre, the same software mistake carries much higher physical and commercial risk.

Even with these issues addressed, there is still the commercial challenge of changing how existing logistics operations work around an autonomous HGV. Removing the driver does not make freight fully automated. Goods still have to be loaded and unloaded, vehicles still need to enter yards and depots, and someone still has to deal with delays. In comparison, a robotaxi alone is able to reliably provide a service once you get the cleaning and servicing sorted in a single fixed depot point.

This systems-integration problem also helps explain why autonomous HGVs may take years, or even decades, to deploy widely across the UK.

Where autonomous lorries are being deployed today

Nevertheless, some start-ups are finding creative routes to commercialisation despite these challenges.

There are about a dozen revenue-generating firms currently serving routes with vehicles that have no driver in the cab. Some still rely on partial remote assistance, while others run fixed routes and handle depot manoeuvres, such as docking at loading bays and positioning vehicles in yards, entirely autonomously. Almost all are start-ups, reflecting the high R&D costs, operational complexity and liability risks involved in launching an autonomous HGV service. Most established logistics firms have so far been reluctant to take on those risks directly.

To give a sense of scale, the UK has about 520,000 registered HGVs and Waymo’s robotaxi fleet, the single largest autonomous vehicle fleet in the world right now, stands at 3,791 cars.

Each of the leading companies in autonomous HGVs shows something different about the maturity of the autonomous freight ecosystem and the niches where the technology is reaching commercial use first.

The largest autonomous lorry provider is Pony.ai, with 210 autonomous lorries and a target of reaching 1,000 by the end of the year. The current fleet generated about $10 million of revenue in Q1 of 2026. China appears to have the largest visible deployments, but it also shows that the industry is not yet fully mature. Pony.ai has experimented with route pooling, with 20 autonomous trucks following one human-driven vehicle at the front. That is a commercially useful model, but it also shows that the sector still has some way to go before large autonomous HGV fleets are operating independently across open road networks.

By comparison, the most promising US start-ups are currently at a scale of dozens of vehicles. There are several firms worth mentioning, including Aurora, Gatik and Kodiak, each representing a different niche in HGV freight.

Aurora has 30 lorries, serving Uber Freight and Hirschbach and operating primarily between Dallas and Houston, which is just a bit shorter than the drive from London to Manchester and a middle-length route by US standards. Aurora aims to reach 200 lorries by the end of this year. It is trying to prove the classic autonomous trucking case: hub-to-hub freight on predictable high-speed routes.

Gatik currently has 41 lorries. Gatik has a deal with PepsiCo and touts a 98% on-time delivery rate in short-haul trips between PepsiCo depots. Its fleet is up 31 vehicles from 10 in January, and it is targeting “hundreds” by the end of this year. Gatik is currently somewhat behind on its expansion schedule, as the plan aimed for 60 lorries by the end of the year. Its niche is middle-mile logistics: repeated journeys between distribution centres, manufacturing sites and retail locations.

Kodiak has sold 28 lorries to Atlas Energy Solutions, which operates them using Kodiak’s technology in the Texan Permian Basin, hauling frac sand on private and leased oilfield roads with no human in the cab. This avoids some of the hardest HGV autonomy problems because speeds are lower and the network is partly controlled, but it also creates different challenges through rough roads, dust, puddles, ditches and routes that change as drilling activity moves.

Each of these firms has a different solution and niche. Pony.ai shows the scale advantage of China’s autonomous freight market, Aurora is focused on long-haul motorway-style freight, Gatik is targeting repeated middle-mile routes, and Kodiak is proving autonomy in harsher but more controlled industrial road networks.

The most notable European HGV player is Sweden’s Einride. The electric, digital and autonomous freight firm recently completed its public listing and has four autonomous vehicles in live customer operations, with plans to reach 20 by the end of the year. Its autonomous work is supported by its larger electric freight business, which operates around 200 electric lorries and sells customers full freight capacity, including charging and drivers for the electric fleets rather than just vehicles. That makes Einride less of a pure autonomy company and more of an attempt to build a full freight platform around electrification, software and autonomy.

Other European players include Volvo Autonomous Solutions, operating in Norway, and Fernride, operating in Estonia, with single-digit HGV fleets.

The overall lesson is that autonomous freight is real, but small and growing. China has the largest visible deployments. The US has several serious public-road and industrial-road experiments. Meanwhile, Europe has the smallest operations, with deployment mostly focused on constrained use cases and broader electric-freight platforms.

This is the first of several autonomous freight pieces. Future pieces will include autonomous vans and spotlights on early-adopting industries.