Many companies are working on self-driving trucks and delivery. As it became clear that robotaxi required a very large investment, teams looked for a more tractable problem. Most have gone after long-haul trucking on the interstate, but one leader has quietly gone after the “middle mile” while others go after long-haul and last-mile.

Gatik is now operating two truck routes with class 4 trucks with no human on board. A few other companies have done demonstration rides with nobody in the truck, but Gatik is doing an actual regular service, moving goods from warehouses to stores for Wal*Mart in Arkansas and Loblaws in Toronto.

Gatik attributes this success to the quality of their own self-driving stack, but also to attacking what they say is the right problem in the middle mile. Long-haul trucking efforts can be found at Waymo, Aurora, tuSimple, Kodiak, Plus, Embark and others. They want to move class-8 semi-trucks distances of 500 miles or more. In the future, that means safer trucking but also cheaper. Existing truck drivers can only work 11 hours/day when solo, while a robot can drive full-time except when refueling. To move cargo that fast, companies hire two drivers to take shifts on the truck while the partner sleeps in the back. It’s grueling and expensive. There’s also a shortage of drivers.

On the other end are last-mile delivery companies like Nuro, Meituan and JD in China. They have small vehicles which drive slowly but must serve a very large number of streets and addresses, including busy, complex streets with pedestrians. There are also the sidewalk delivery companies which go after a much simpler problem but can only carry more limited amounts of cargo over short distances. The leader, Starship (I’m a shareholder) has now done over 4 million commercial, autonomous deliveries.

Gatiks’ sweet spot is medium sized trucks over a limited set of short routes, with trips in the range of 20 to 30 miles. The single short fixed routes are easier to develop for and certify. You don’t scale to the whole road network so scaling is easier. Their business is entirely B2B, right now moving goods from warehouses to stores and other warehouses to enable a just-in-time supply chain.

For Loblaws grocery, they enable online ordering with a 3 hour pickup time. Consumers can order anything in the warehouse, not just what is stocked in their local store. The warehouse orders are batched into a small truck and go to the store to be combined with locally stocked items and ready for the customer to either pick up or get last-mile delivery. Small trucks are more efficient and can be run more often when you don’t have to have extra drivers. A lot of stores use full sized Class 8 trucks for middle-mile which can often be quite wasteful. The routes are all semi-urban, aterials and highway, which present a simpler driving problem.

Middle-mile is also easier from a regulatory standpoint, as you rarely or never cross jurisdictional boundaries, unlike long-haul trucks which cross state borders.

This advantage is similar to the advantage self-driving can bring to transit. For public transit agencies, who get grants to pay for infrastructure, driver salaries can be the largest expense. While riders would prefer to see 10 vans in an hour instead of 2 buses, there’s no way agencies could hire that many drivers. When there are no drivers, the smaller vehicles work fine, and because they attract more riders, they actually will be more energy efficient than the large buses on a per-passenger basis. Gatik allow them to dispatch smaller trucks more frequently, providing better service and increasing demand.

While Gatik currently does simple single routes with no safety driver, and more complex routes with a safety driver, their plan is to open up a service where one warehouse will serve 34 different Sam’s Club stores in its area. While validating 34 routes is more work, it’s easier than trying to make a robotaxi or general delivery service. Gatik is small compared to companies like Waymo and tuSimple (which just laid off 25% of its staff.) Gatik has about 100 employees in Silicon Valley and Toronto and has enough funding to last 3 years in turbulent times.

Like all delivery offerings, with no passengers there is nobody in a vehicle to be hurt in an accident. Even so, the vehicles are still much larger than cars and must very clearly avoid collisions with cars or vulnerable road users. The problem is simple and the stakes are high, but the reward is mainly cheaper and more flexible shipping — and getting around the driver shortage. Drivers for middle mile are more available because the job doesn’t demand leaving family for long periods of time. So it’s not a perfect sweet spot but easier than some of the other tasks.