Chinas Robo-Taxis Are Coming to Seoul. Seoul Should Be Worried.
Pony.ai plans to deploy 200 self-driving vehicles in South Korea by 2028 through a local partner — a move that exploits Korea's regulatory straitjacket and data-starved industry while China's robotaxi firms pivot west after losing the US market.
The Quiet Conquest of Seoul’s Streets
Pony.ai is planning to roll 200 self-driving vehicles onto South Korean roads by 2028. That number is small in absolute terms but enormous relative to what currently drives around Seoul — 19, operating only at night in a single district. The contrast is not accidental. It is the clearest possible statement about where the global robotaxi race stands.
China’s autonomous-driving firms entered overseas expansion mode this year, and South Korea is now one of their targets. The United States effectively closed its door after the Biden administration banned Chinese and Russian software and components in vehicles sold from 2027 models onward. So Baidu looked west, testing in London. WeRide signed up for France, then added Spain. Pony.ai chose Korea — a country whose domestic robotic taxi fleet is so constrained that Chinese operators can drive around it like a training simulation before scaling up elsewhere.
The strategy behind Pony.ai’s entry is worth studying. Rather than arriving with a fleet and facing immediate backlash, the company took a two-year detour through a local partnership. In 2024, Pony.ai transferred its autonomous-driving kit to Future Link, a Korean firm. Future Link mounted the hardware on Hyundai Kona vehicles and ran 10 cars under a temporary operating permit in Gangnam, accumulating 87,000 kilometers of real-world data. Domestic car, domestic company, minimal political friction — and a foothold in Korean roads.
That was phase one. Phase two is the current agreement signed on August 28: 200 Chinese-made robotaxi vehicles delivered by 2028, built around Pony.ai’s seventh-generation platform. The sensor suite jumps from 16 units on the Kona conversions to 34 on the new fleet, delivering 360-degree coverage and a 650-meter detection range. Phase three — still unannounced in equity terms — will be a joint venture, with Pony.ai supplying technology and vehicles while Future Link handles operations, data management, and local adaptation.
Why Korea Is the Easy Target
The reason Pony.ai chose Korea has less to do with opportunity and more to do with vulnerability. South Korea’s autonomous-driving sector is functionally frozen in place.
Domestic operator AutoNavi A2U had accumulated just 690,000 kilometers of driving data as of last August. By comparison, Google’s Waymo has logged 160 million kilometers. Baidu’s Apollo Go has 240 million. The gap is not incremental — it is two orders of magnitude. Data is the fuel of autonomous driving, and Korean firms are running on fumes.
The root cause is regulation. The Korean government permits robotaxi operations only within 55 designated pilot zones across the country, and only along fixed routes. The Passenger Vehicle Transport Service Act further restricts commercial service: any private operator needs a taxi license to run paid rides. The combination has produced a market where 19 robotaxis can operate at night in one district of Seoul, and no Korean company has broken out of that containment.
Meanwhile, China’s big players are running thousands of vehicles 24 hours a day in cities like Wuhan, Beijing, and Guangzhou. By the time Pony.ai arrives in Seoul, its drivers will already have logged millions more kilometers than any Korean operator. The technology gap is not theoretical — it is operational.
The Regulatory Hurdles Are Real but Manageable
Pony.ai will face certification gates before hitting the road in quantity. The Ministry of Land, Infrastructure and Transport and the Korea Transportation Safety Authority must validate crash safety and braking performance. The Ministry of Environment will test range per charge. The autonomy system itself requires a separate safety certification. Commercial service needs permission from the relevant provincial or metropolitan governor’s office.
On paper, these barriers apply equally to all manufacturers regardless of origin. In practice, the political calculus is different. Vehicles equipped with LiDAR scan Korean roads in high resolution — a detail that could trigger security concerns and slow approval timelines. Pony.ai’s CEO James Feng has publicly committed to storing all collected data locally and anonymizing faces and license plates, which should mitigate the most obvious objections.
But the timing is telling. Korea is not yet ready to give its own companies the regulatory room to compete. And Pony.ai is arriving precisely when those constraints are most evident.
Who Wins, Who Loses
The immediate winners are Pony.ai and, to a lesser extent, Hyundai Motor, which gets a high-visibility deployment partner and a chance to move autonomous-driving hardware on local roads. Future Link benefits from a technology transfer and a pathway to scale that Korean law has denied its competitors.
The losers are Korean autonomous-driving firms that have been trying to grow inside a regulatory box. If Pony.ai operates commercially in Seoul while domestic rivals remain confined to 19 cars at night, the narrative will shift fast — and investors, partners, and talent will follow the winner. There is also a second-order risk: Korean consumers who see a Chinese robotaxi operate reliably may lose faith in local providers before those providers have had a chance to improve.
What This Means Beyond Korea
Pony.ai’s Korean play is a test case for a broader pattern. Chinese robotaxi firms are now operating on three continents — North America is blocked by policy, Europe is open but nascent, and East Asia is wide open. Korea is the most symbolically significant entry point because it is a technology-exporting nation that has failed to export autonomy. If Chinese software can run on Korean roads while Korean software cannot leave the country, the implication extends far beyond transport.
The global robotaxi race is no longer about who builds the best car. It is about who accumulates the most data fastest — and who faces the fewest regulatory barriers to doing so. Pony.ai’s Korea strategy is elegant precisely because it exploits both sides of that equation at once.