Why autonomous trucking is finally back, three years after the hype collapsed

Three things, mostly.

The first is hardware cost. Sensor stacks for highway-grade autonomy are roughly half what they cost in 2022. Compute costs per truck are similar.

The second is route economics. Limiting autonomy to specific corridors means the cost of mapping, monitoring, and operating-domain validation is amortised over far more vehicle-miles than the original any-route model could support.

The third is customer demand. The trucking industry is short of long-haul drivers. Autonomous augmentation on the long-haul middle of a route, paired with human drivers on the more complex ends, is an easier sell to fleet operators than total replacement.

What is likely next

Expect the next 18 months to feature gradual route expansion, not dramatic announcements. Expect the survivors to add specific corridors one at a time, with a quiet emphasis on incident-free operating data.

Expect at least one of the larger automotive OEMs to acquire a survivor before 2028. Vertically integrated truck plus autonomy plus fleet ownership is the natural endgame, and the time to consolidate is when the technology works but is not yet at full scale.

Autonomous trucking spent 2022 promising too much. It is spending 2026 quietly delivering less. That is the right direction.

Zoox begins revenue service in Las Vegas with a vehicle nobody expected

Amazon's autonomous vehicle subsidiary has launched its first commercial service outside San Francisco.

Helena Frost · 4 May · 3 min