Tesla’s plan to turn autonomous ride-hailing into a major business is facing growing competition from China. XPeng, one of the country’s best-known electric-vehicle manufacturers, has begun mass-producing a purpose-built robotaxi and intends to launch pilot services before moving toward fully driverless operation.
The challenge matters because XPeng is following a strategy similar to Tesla’s. It is developing the vehicle, artificial-intelligence software, computing hardware and autonomous-driving system largely within the same company. It also plans to use that technology beyond private cars, expanding it into robotaxis, humanoid robots and even flying vehicles.
Tesla currently operates its Robotaxi service in limited areas of Austin, Dallas and Houston in Texas, as well as Miami, Florida. XPeng’s service remains focused on China, so the two companies are not yet competing for the same passengers. However, their overlapping technology strategies and international ambitions could eventually make them direct rivals.
XPeng Has Started Producing Its First Robotaxi
XPeng began mass production of its first dedicated robotaxi at its Guangzhou headquarters in May 2026.
The vehicle is based on the company’s large GX electric platform and is designed as a factory-built autonomous vehicle rather than a conventional car modified after production. XPeng describes it as a production-ready robotaxi developed using its own technologies, including proprietary artificial-intelligence chips and autonomous-driving software.
Pilot operations are expected to begin during the second half of 2026. The company aims to remove the onboard safety driver and begin fully driverless operation in early 2027, although commercial deployment will still depend on testing results and regulatory approval.
XPeng President Brian Gu said the company could produce anywhere from several hundred to several thousand robotaxis during the first 12 to 18 months. That would still represent a relatively small fleet, but it could provide the real-world mileage needed to improve the system before a larger rollout.
XPeng had already announced that its robotaxi program, VLA 2.0 autonomous-driving technology and next-generation IRON humanoid robot were scheduled to enter mass production in 2026. The latest manufacturing milestone shows that the robotaxi project has moved beyond a presentation or experimental prototype.
Its Technology Resembles Tesla’s Approach
Most major robotaxi operators use several types of sensors, typically combining cameras with radar and lidar. Lidar uses laser pulses to create a detailed three-dimensional picture of the vehicle’s surroundings.
XPeng’s robotaxi reportedly takes a different approach. Like Tesla, it relies primarily on cameras and an end-to-end artificial-intelligence model instead of lidar and highly detailed pre-mapped routes.
The system is based on XPeng’s VLA 2.0 technology. VLA stands for vision-language-action, describing an AI architecture intended to interpret visual information, understand driving situations and choose an appropriate action.
XPeng has also developed its own Turing AI chips to provide the computing power required inside the vehicle. This vertical integration could help the company control costs and update the complete autonomous-driving stack more quickly.
Tesla likewise relies on a camera-centered system and neural-network software rather than the sensor-heavy approach used by companies such as Waymo. That shared philosophy makes XPeng a particularly interesting competitor because it is attempting to prove that scalable autonomy can be achieved with less expensive hardware.
The approach could lower the cost of each robotaxi, but it also creates difficult safety questions. Cameras may struggle when lenses are blocked, visibility is poor or lighting conditions are unusual. Removing lidar also eliminates a separate type of sensor that could provide additional information when cameras face limitations.
Neither company has yet demonstrated that a camera-dominant system can operate safely and commercially at enormous scale across every type of road and weather condition.
Tesla Has Already Launched, but Its Footprint Remains Limited
Tesla introduced its Robotaxi service in Austin in June 2025 and has since expanded into several additional U.S. cities.
According to Tesla’s official Robotaxi information, rides are currently offered in limited areas of Miami, Austin, Dallas and Houston. Passengers book through a dedicated application, and availability depends on whether the requested destination falls within a supported service zone.
This gives Tesla a practical advantage over XPeng: paying customers are already taking rides.
However, Tesla’s network is still small compared with the nationwide service its long-term strategy envisions. The company must demonstrate that its software can handle a much wider range of streets, weather, emergency situations and local traffic rules before robotaxis become a major source of revenue.
Tesla is also developing the Cybercab, a two-passenger electric vehicle without traditional steering controls or pedals. The purpose-built model is intended to reduce vehicle and operating costs when deployed in large autonomous fleets.
Until the Cybercab becomes widely available, Tesla can use existing vehicles such as the Model Y in its Robotaxi network. That gives the company flexibility, but a conventional passenger vehicle may cost more to clean, maintain and operate than a vehicle designed specifically for continuous ride-hailing.
XPeng’s Large Robotaxi Could Serve More Passengers
The physical design of XPeng’s vehicle may distinguish it from Tesla’s Cybercab.
Tesla designed the Cybercab as a compact two-seater. That configuration may work well for individual passengers and couples, which account for many ride-hailing trips. It may be less practical for families, larger groups, people carrying substantial luggage or customers requiring more accessible transportation.
XPeng’s GX-based robotaxi is a large electric SUV designed around passenger comfort. Reports describe rear privacy glass, reclining zero-gravity seats and voice-controlled entertainment features.
A larger vehicle will generally require more energy and may cost more to manufacture. It could nevertheless generate more revenue per trip if it can transport additional passengers or serve premium and airport routes.
This highlights an unresolved question in the robotaxi industry. The lowest-cost vehicle may not always create the best service. Fleet operators must balance vehicle price against passenger capacity, cleaning time, accessibility, energy use and expected demand.
Tesla appears focused on an inexpensive, highly optimized two-passenger vehicle. XPeng is initially emphasizing a more spacious passenger experience.
China Offers a Valuable Testing Environment
China may give XPeng an important advantage because its cities already support numerous autonomous-vehicle trials.
Baidu’s Apollo Go, Pony.ai and WeRide have accumulated extensive robotaxi experience in Chinese cities, creating a more mature regulatory and technical environment than many countries. Consumers in several urban areas can already request autonomous rides through established platforms.
That means XPeng is not entering an empty market. It will face strong domestic competition, but it can also benefit from clearer operating rules, available infrastructure and a population that has already been introduced to driverless taxis.
China’s dense urban environments provide valuable training conditions. Autonomous vehicles must handle scooters, bicycles, pedestrians, delivery vehicles, crowded intersections and unpredictable traffic behavior.
Successful operation in those conditions could strengthen XPeng’s software. Failure could expose weaknesses much faster than controlled testing on simpler roads.
Low ride prices may also encourage adoption. Research involving robotaxi passengers in Wuhan found that cost sensitivity and willingness to use the technology were among the strongest predictors of actual adoption.
XPeng Is Already Expanding Beyond China
XPeng was founded in 2014 and already sells electric vehicles outside its domestic market. The company has entered numerous European and Asian countries while maintaining research operations in locations including Silicon Valley.
Its international presence does not mean robotaxis will immediately follow its consumer vehicles. Autonomous-driving rules vary significantly between countries, and software trained in China may require extensive adaptation before operating elsewhere.
Nevertheless, XPeng has indicated that it could introduce autonomous-driving technology in Europe as early as 2027. The company is also working with Volkswagen, which may give it access to additional engineering, manufacturing and distribution experience.
XPeng could choose several expansion models. It may operate its own ride-hailing service, license the technology to another mobility company or sell completed robotaxis to third-party fleet operators.
The third option could allow it to enter markets without building a consumer-facing service from the beginning. Tesla, by comparison, is attempting to control the vehicle, software, application and ride-hailing network.
Regulation May Matter More Than Manufacturing Speed
Producing a robotaxi does not automatically give a company permission to operate one without a driver.
Regulators need evidence that an autonomous system can identify hazards, obey traffic rules, respond to emergency vehicles and reach a safe condition when something fails. Authorities may also require remote-support systems, incident reports, cybersecurity protections and adequate insurance.
Unusual situations remain one of the industry’s greatest challenges. Construction zones, hand signals from police officers, damaged traffic lights, heavy rain and emergency scenes can confuse systems that perform well during ordinary driving.
Robotaxi operators have already faced investigations and recalls involving vehicles entering restricted areas or responding incorrectly to unexpected road conditions.
XPeng’s target of removing safety drivers in 2027 should therefore be viewed as an objective rather than a guaranteed launch date. Tesla’s expansion will face similar regulatory limits as it moves into additional U.S. cities.
Public trust will also influence adoption. A small number of widely shared failures can damage confidence even when a fleet’s overall safety statistics appear favorable.
The Competition Extends Beyond Robotaxis
XPeng and Tesla are increasingly beginning to resemble broader artificial-intelligence and robotics companies rather than traditional automakers.
XPeng plans to commercialize its IRON humanoid robot and has discussed producing more than 1,000 robots per month by the end of 2026. It is initially expected to place them in customer-service roles before pursuing international expansion.
The company’s AeroHT division is also developing modular flying vehicles. The project reflects XPeng’s strategy of applying batteries, electric motors, AI chips and autonomous-control software across multiple forms of transportation.
Tesla is following a comparable path through its Optimus humanoid robot, autonomous vehicles, energy systems and AI development.
The overlap matters because advances in one project can support another. A better vision model may improve both a humanoid robot and a driverless vehicle. Larger chip orders can reduce hardware costs, while additional real-world data may strengthen future AI systems.
XPeng Is a Credible Threat, but Tesla Is Not Displaced Yet
XPeng has several qualities that make it a plausible future competitor to Tesla: large-scale EV manufacturing, proprietary chips, end-to-end AI development and access to China’s fast-moving autonomous-vehicle market.
It has also moved quickly from announcing a robotaxi to beginning production. If pilot operations progress successfully and safety drivers are removed in 2027, XPeng could become one of the first traditional EV manufacturers to operate a vertically integrated driverless taxi service at meaningful scale.
However, production is only the beginning. XPeng still needs to prove that its vehicles can operate safely, earn regulatory approval and generate sustainable revenue. It must also compete with experienced Chinese operators before challenging Tesla internationally.
Tesla retains advantages in brand recognition, software data, U.S. manufacturing and an active commercial Robotaxi service. Its greatest risk is that competitors such as XPeng may improve and expand faster than expected.
The robotaxi race will not be decided by which company produces the most impressive prototype. It will be decided by safety, cost, fleet uptime, regulatory acceptance and the ability to complete millions of uneventful rides.
XPeng has now entered that race with a real production vehicle, making it much harder for Tesla to assume that its greatest autonomous-driving competitors will come only from the United States.
