Pony.ai showed its fourth-generation autonomous truck in Guangzhou this month and said it plans to operate between 500 and 1,000 self-driving trucks within two to three years. The truck unit started in 2018 from near zero, when a small team bought a Dongfeng rig and rewired throttle, brake and steering themselves. Eight years on, the product is moving from demo to fleet.
Vice president He Xing, who leads the truck business, rejects the idea of a pivot. Pony.ai simply waited, he says, until the stack was mature enough to mass-produce, the cost model was clear, and the firm knew how to run a fleet. The Robotaxi base carries the truck: light trucks share over 95 per cent of code with Robotaxi, and heavy trucks still reuse over 80 per cent of code and over 90 per cent of hardware.
Where the trucks go first
Scenes are trunk lines, bulk-commodity routes in the northwest, and port horizontal transport. In ports, where autonomous driving is harder than it looks because manned and unmanned vehicles mix, Pony.ai’s city-driving experience lifted single-loop efficiency by more than 30 per cent over rival schemes in a major port group’s tests.
The cost case is ton-kilometre: northwest routes face driver shortages through harsh winters, and ports run 7×24 against a three-shift driver need. Pony.ai offers two models. ADaaS hands a “virtual driver” to partners who own and run the vehicles; TaaS has Pony.ai or partners manage the fleet and sell transport. Long term, He says, Pony.ai exits the chain and becomes a pure technology supplier.
Robotaxi is already profitable in two cities. The next three years are meant to prove autonomous trucks can do the same, and become a real transport business.
Editor’s note: This is an adapted translation of the original LeiFeng Network report. It has been trimmed and restructured for readability for an international business audience.