200 Chinese Robotaxis Are Coming to Seoul by 2028

The deal

Chinese autonomous driving company Pony.ai and Korean firm FutureLink signed a strategic agreement in Seoul on August 28. The headline number is 200 robotaxis destined for Korean roads. The plan is to convert technical validation work already conducted in Seoul's Gangnam autonomous driving pilot zone into a commercial service, launching around 2028 and then expanding beyond Seoul to other Korean cities.

The rollout is staged. An initial batch of roughly 10 vehicles will go through Korean certification; if that clears, the remaining 190 follow. The vehicles are reported to be BAIC-based models fitted with Pony.ai's seventh-generation autonomous driving system. The division of labor is fairly clean: Pony.ai supplies the tech and vehicles, while FutureLink handles Korean certification, market entry, fleet and service operations, data management, and localization. Both companies emphasized that driving data collected in Korea would be stored domestically.

The target is Level 4 — driverless operation within a defined operational design domain. The 2028 date should be read as a goal contingent on certification and regulatory groundwork proceeding smoothly, not a firm launch date.

Why now

Pony.ai, a Nasdaq-listed company, already runs driverless robotaxi services in Chinese cities including Guangzhou and Beijing, and has been signing partnerships across the Middle East, Southeast Asia, and Europe. Korea fits that pattern. In autonomous driving, cumulative mileage and edge-case data are the moat, so geographic expansion is the strategy in itself.

Korea is an appealing proving ground. Its urban roads are narrow and dense, with heavy mixing of pedestrians, scooters, and delivery motorcycles — genuinely hard conditions. At the same time, telecom infrastructure and HD mapping coverage are world-class, and the government's pilot zone framework has lowered the barrier to real-world testing. A system validated in Seoul travels well as a reference for other Asian markets.

How Korea's own players compare

Context matters here. Seoul has run late-night autonomous buses and robotaxi pilots in Gangnam for a few years, but almost all of it involves a safety operator on board and tightly limited routes. Hyundai Motor Group has built robotaxi capability largely through Motional, its US-based venture, and several Korean startups are scaling pilots — but no fully driverless commercial service exists in Korea today.

Meanwhile Waymo operates paid driverless rides across multiple US cities, and in China, Baidu's Apollo Go, Pony.ai, and WeRide have all moved into commercial operation. The global race has effectively consolidated into a US-China duopoly. What this agreement highlights is the possibility of Korea — a country with a top-five global automaker — becoming a stage for imported autonomy rather than an exporter of it. For domestic carmakers and suppliers, that is an uncomfortable but useful signal.

The data and security question

Predictably, the announcement immediately drew concern about Chinese technology operating on Korean streets. A robotaxi is essentially a mobile sensor platform, continuously capturing lidar point clouds and camera footage of roads, buildings, and people, alongside precise positioning data and passenger movement patterns. The partners' emphasis on domestic data storage reads as a preemptive response to exactly this criticism.

Storage location and access rights are separate issues, though. Data held on Korean servers can still be reached remotely for model training or technical support, and the practical question is how regulators verify that. Both the EU and the US have tightened scrutiny of Chinese-made connected vehicles for the same reason. In other words, this project's fate will likely be decided at the certification and security-review stage rather than on technical merit.

What changes for riders

Nothing changes next year. The 2028 timeline depends on certification, operating permits, and unresolved questions around insurance and liability in driverless crashes all lining up. If it does proceed, riders could see three shifts.

First, more late-night mobility options — a service with no driver-supply constraints and no ride refusals has real value in a city where late-night taxi shortages are a recurring complaint. Second, potential movement in taxi fare structures, though overseas experience suggests early robotaxi fleets are not cheap: sensor hardware and remote-monitoring staff keep unit costs high. Third, friction with the incumbent taxi industry. Two hundred vehicles is trivial next to Seoul's total taxi fleet, but the symbolism and the expansion trajectory are what matter politically.

Bottom line

This is an agreement, not a delivered fleet, and a substantial certification hurdle sits between the announcement and any paying passenger. Its significance is that it forces a question into the open: does Korea want to build autonomous driving technology, or import it? How regulators balance approval speed, data sovereignty, and support for domestic developers over the next two to three years will determine whether these 200 vehicles ever pick up a rider.