Xpeng targets overseas robotaxi trials in 2027, with cities undisclosed

Xpeng plans to begin robotaxi trials in selected cities outside China in 2027, vice-chairman and president Brian Gu told CNA in an interview in Singapore reported on October 9. The proposed pilots would extend its autonomous-driving testing overseas, but Gu did not name destinations or announce when passengers could use a commercial service.

“Starting next year, we hope to launch pilot testing in select global cities,” Gu told CNA’s Julie Yoo on the sidelines of the Milken Institute Asia Summit. His wording sets out an intended testing timetable; the report does not establish that overseas trials have been authorised or begun.

Xpeng’s Guangzhou pilot and overseas ambitions

CNA reported that Xpeng is already testing its Yoyo robotaxi in Guangzhou, southern China, with close to 100 vehicles involved. That figure describes the Chinese pilot. The interview did not establish the size of any overseas fleet, local partners, exact trial dates or fares.

Gu also offered a broader forecast for the industry: “Robotaxis are not that far from everyday life. As soon as probably two to three years later, you could see larger-scale commercial deployment around the world.” Relative to the October 2026 report, that points approximately to 2028–2029.

Driver assistance differs from driverless operation

Gu said Xpeng’s driver-assistance systems were used for nearly 60 per cent of the distance travelled by its vehicles. The statistic describes use of assistance systems, rather than mileage completed by driverless robotaxis. CNA’s account does not provide a methodology or independent validation of the figure.

Fully autonomous robotaxis face a higher standard, Gu acknowledged: they must safely handle complex and unexpected situations without a human driver. The report provides no independently validated safety results or intervention rates for the Guangzhou trial, leaving its performance beyond the reported fleet size unestablished.

The details appear in CNA’s interview report. Its publication date establishes when the plans were reported; the exact interview date is not stated.

Singapore’s separate Waymo timetable

Singapore offers a separate example of the stages involved in bringing an autonomous-vehicle operator into a new market. It is not a confirmed Xpeng destination: Gu’s presence there for the interview does not identify it as one of the cities under consideration.

In a September 18 announcement, Singapore’s Land Transport Authority said Waymo would join ComfortDelGro and Grab as an autonomous-vehicle operator. Vehicles were expected to arrive in the following few months, with initial manual driving in 2027 and a target of commercial ride-hailing in 2028.

Those are announced milestones, rather than confirmation that commercial operations have started. LTA said it would work with Waymo on safety and regulatory requirements, describing its entry as supporting more transport options for commuters and a competitive autonomous-vehicle industry.

Safety assessments and support for drivers

LTA’s autonomous-vehicle guidance requires safety assessments at the CETRAN test centre before vehicles can be deployed on Singapore’s roads. Operators also need insurance covering third-party liability and property damage, and vehicles must carry a data recorder. Trial applicants must identify their proposed operating area and apply to LTA.

The framework separates deployable and developmental solutions. The deployable track requires realistic deployment plans and a record of similar use cases involving multiple vehicles over at least one year. These Singapore requirements provide context for assessing deployment plans; they do not establish Xpeng’s approvals or obligations in its undisclosed target cities.

In its Waymo announcement, LTA also said deployment would not outpace the government’s ability to retrain and support affected drivers. It cited a July 2026 manpower transition package and help for willing drivers to acquire skills and move into new roles. That is a policy commitment, not evidence of measured job losses.