Gasgoo Munich- "I paid 400,000 yuan for this car, yet its smart driving is worse than a 100,000-yuan family runabout."
This is the reality for hundreds of European owners of Chinese cars. Shortly after deliveries, the driver-assistance system started glitching—sometimes slamming on the brakes at high speeds for obstacles that didn't exist. The manufacturer tried repeatedly to fix it but never produced a software patch compliant with the EU's WVTA certification. In the end, the company permanently disabled multiple driver-assistance features via an OTA update and offered owners cash compensation.
Such struggles are becoming common as Chinese smart cars expand abroad. Liu Xiaobei, CTO of Che Ling Tech, surveyed the connectivity of Chinese exports and found that roughly 70% are effectively offline. Of the remaining 30% that are connected, only about 10% of users continue paying for data services in the second year.
From disabled driver-assistance to long-term disconnection after deliveries, the ordeal facing Chinese smart cars overseas has extended well beyond sales and after-sales service.
Is the Car "Frozen" the Moment It's Delivered?
In the domestic market, a key selling point of smart cars is that they stay "fresh." After deliveries, automakers push new features and fixes via over-the-air (OTA) updates, constantly improving the experience. The longer you own the car, the more it evolves.
But overseas, this model runs into constraints from data flows and local regulations. Zhang Xiaodong, deputy director of Geely Automobile Group's Technical Management Center, noted at a recent conference that restrictions on cross-border data flows directly impact OTA updates for vehicles abroad.
Software upgrades are also critical for continuous optimization on real roads. Driver-assistance systems face different road environments, traffic participants, and driving habits every day. This requires R&D teams to spot issues based on actual vehicle operation and fine-tune the system through updates.
Data shows a single smart-connected vehicle generates about 10TB of data daily—5 to 10 times that of a traditional internal combustion engine car. This data covers road conditions, driving behavior, and vehicle status, serving as a vital basis for automakers to analyze performance and refine algorithms.
Overseas vehicles are equally important sources of real-world data. The longer they operate locally, the more automakers need to understand the actual roads and user habits. If that data cannot flow back into R&D and product optimization, it becomes difficult to adjust software based on actual performance abroad. Software iteration after deliveries loses a crucial feedback loop.
Image Source: Huaban.com
Software services also suffer from connectivity issues. Domestic automakers have integrated online functions, software services, and continuous updates into user operations, but some features require a constant connection. If a vehicle stays offline for long periods, these online functions fail to deliver, making it hard for automakers to build stable user operations or revenue streams around them.
Functional differences hit users directly. One blogger noted that Chinese EVs bought overseas often lack CarPlay, forcing drivers to use third-party apps for navigation. But once the car starts moving, these casting functions are often restricted.
Driver-assistance features face similar hurdles. The same system entering different countries may unlock different functions depending on local regulations, product certification, and road conditions.
In other words, the high-level driver-assistance capabilities advertised at home might only be partially available in some overseas markets. These gaps in the intelligent experience directly affect user reviews and, ultimately, the long-term brand building and sustained competitiveness of Chinese automakers.
Zhang Yanhua, president of the China Automotive Maintenance and Repair Trade Association, pointed out that Chinese brands have already crossed the threshold of product credibility. The next competitive frontier extends to after-sales service, repair networks, insurance support, and full lifecycle management. The new challenges brought by smart car exports are thus stretching well beyond deliveries into a much longer usage cycle.
This Isn't a Capacity Issue, It's a Regulatory Issue
In recent years, a key path for Chinese automakers going global has been localization: building factories, expanding channels, and following supply chains. But building a factory doesn't solve the "offline" problem.
Moreover, smart cars must navigate data flow rules in every country. The EU's GDPR imposes strict requirements on processing and cross-border transfers of personal data—vehicle operation data, in-cabin camera footage, and street imagery can all involve personal information. Australia's Privacy Act and regulations in some Southeast Asian nations also impose requirements on data storage and transmission.
Restrictions in South Korea are even more complex. The country not only limits data exports but also sets thresholds on data access frequency and volume.
These regulations directly impact corporate data processing workflows. Chu Huabin, general manager of Quantiphi (China) Information Technology, gave an example: a smart driving team conducting road tests in Southeast Asia might generate hundreds of gigabytes of street imagery, point cloud data, and algorithm decision logs daily.
Image Source: Leapmotor
R&D teams need this data to analyze road environments and debug algorithms, but faces and license plates in the imagery involve personal information. Raw data cannot be exported directly; it must be anonymized locally, and license plates require tokenization.
Data processing is therefore scattered across different markets. Chu noted that some companies, deterred by the complexity and compliance risks, choose to process data locally. This makes it harder for domestic R&D teams to access overseas data, slowing algorithm iteration. And with every new market, companies must relearn local rules and configure data processes accordingly.
Cross-border collaboration challenges extend to physical parts. Wen Hua, deputy general manager of Dongfeng Liuzhou Motor's import-export division, noted that faulty parts from overseas repairs usually need to be shipped back to China for analysis. But some markets control the import of used parts, requiring specific paperwork. If problematic parts can't return quickly, domestic R&D teams can't analyze issues promptly, dragging out quality improvement cycles.
These issues aren't unique to Chinese firms. American EV maker Rivian's official policy states its service coverage primarily spans the U.S. and Canada; outside these areas, connectivity services and remote diagnostics may be limited.
But it's worth noting that intelligent technology is a core competitive edge Chinese automakers have built in recent years, and their overseas business is expanding. As vehicles enter more countries, adapting to local data regulations is essential. The ability to leverage smart features is critical to maintaining product competitiveness.
Europe deserves particular attention. The region has strict requirements for user data privacy. Data shows that in the third quarter of 2026, Chinese brands captured a 12.7% share of the EU market, surpassing Japanese automakers for the first time. A larger market scale means more vehicles, users, and road data are entering the region. In this context, data compliance and cross-border collaboration will be a mandatory lesson for Chinese automakers going global.
Automakers Can Build a Factory, But Building an R&D System Is Harder
If overseas data cannot be transmitted back, automakers may need to deploy computing power and R&D facilities locally. But that investment is far from trivial.
Lu You, product and market general manager at neueHCT, crunched the numbers: a single overseas data center of comparable scale costs hundreds of millions of yuan. Another industry insider mentioned that if data cannot be sent home, companies must redeploy computing power and models abroad, with initial investment for a single market potentially reaching tens of millions.
Currently, domestic automakers face squeezed R&D cycles and falling vehicle prices. They must control development costs while maintaining iteration speed. If a new R&D facility is required for every overseas market, the upfront investment is hard to recoup through the scale of a single market.
AI generated; Source: Doubao
Beyond infrastructure, inconsistent data standards add to the difficulty of cross-border collaboration. Hou Weiqiang, sales VP for LexisNexis Risk Solutions in China, cited an example: a single car has a R&D code at the OEM, an industry model code in insurance, and yet another code at parts suppliers. Different industries maintain their own standards without a unified mapping.
This information gap affects insurance pricing for Chinese cars overseas. Data shows the average premium for UK drivers is 646 pounds, but for Chinese exports to the UK, it jumps to 901 pounds—about 40% higher. Hou believes a key reason is that local underwriters lack sufficient information about Chinese exported vehicles.
Insurers need to know vehicle configuration to assess accident risk and repair costs. If they cannot accurately identify specs or lack enough data to evaluate new technologies, they must adopt cautious pricing. The added smart features in Chinese cars might therefore be factored in as higher repair risks.
In China, policy is already providing clear compliance paths. In early 2026, the Ministry of Industry and Information Technology and seven other departments issued the "Guidelines on Security of Car Data Outbound (2026 Edition)," clarifying three paths: security assessment, standard contracts, and personal information protection certification, along with nine exemption scenarios.
Among them, critical data related to security vulnerabilities, incidents, and OTA software source code are exempt from security assessment declarations.
The Guidelines also set varying requirements based on the volume of personal data exported. When filing, companies only need to provide characteristic information like data type and scale, not the data itself.
These rules provide a clear operational basis for Chinese companies handling car data exports, reducing compliance uncertainty for some businesses. However, companies still need to establish data management systems, allocate necessary technology and personnel, and coordinate R&D, operations, and service teams across different countries.
The Second Half of Going Global Is About the Invisible Road
To build this cross-border data system, Chinese automakers currently have at least three directions to pursue.
The first path is turning individual compliance challenges into platform-based public services.
Chu Huabin proposed the idea of "Data Cross-Border as a Service," using a unified entry point and end-to-end contracts to lower the cost of studying regulations and building compliance processes market by market. He noted that his team's platform has built eight trusted data subspaces for the automotive industry in Hainan and identified 86 automotive data scenarios.
The value of such platforms lies in centralizing scattered compliance services. When entering new markets, companies can use public platforms to understand local requirements, reducing the cost of duplicate construction and trial-and-error.
The second path involves local R&D. Geely has integrated European engineering resources to establish a European tech company, largely to meet local data and cybersecurity requirements. XPENG's Munich R&D center aims to localize development for smart driving and smart cabins.
Cloud providers are also offering regional deployment solutions. Alibaba Cloud proposes a "Regional Center + Country Station" model. GAC Group deployed its Middle East regional hub node on Alibaba Cloud and built a telematics system in Saudi Arabia to advance local data storage and compliance certification. This approach allows automakers to align R&D and data facilities with regional business layouts, gradually building local data processing and R&D capabilities.
Image Source: BYD
The third path is unifying data standards across the supply chain.
Hou Weiqiang argues that the industry needs data hubs and unified coding rules so OEMs, insurers, and parts suppliers can identify and use the same vehicle information. Only by establishing accurate mappings between R&D codes, sales names, insurance model codes, and parts codes can parties exchange information smoothly.
This also affects overseas insurance business. The more accurate information insurers have about vehicle configuration and repair risks, the easier it is to judge a car's actual risk. This helps reduce pricing deviations caused by information gaps when Chinese vehicles are insured abroad.
Each path has its focus. Public platforms help cut compliance costs, overseas infrastructure supports local R&D, and unified standards enable smoother collaboration across the supply chain. But they all point to one thing: Chinese automakers' globalization must extend from products and capacity to R&D, data, and service systems.
After all, every car entering a new market must "retake the exam" based on local conditions. This involves product adaptation, but also regulations and services. Only by building sustained operational capabilities locally can automakers truly adapt their products to the market.
Speaking at the conference, Sun Xiaohong discussed goals for the "15th Five-Year Plan" regarding automotive exports: standing firm and bringing value back. Automakers must gain a foothold overseas while ensuring that experience and data accumulated abroad flow back into the R&D system under compliance conditions to support product improvement. This is a capability Chinese automotive globalization must build next.
In the past, automakers competed on screens, computing power, and driver-assistance features. Now, as these products go global, competition has extended to data collection, compliance processing, and cross-border collaboration. Products can be shipped overseas quickly, but R&D and service systems require long-term construction.
Once wheels cross the border, data must find a compliant path to follow.
