Gasgoo Munich-On the streets of Paris, Chinese autonomous driving technology is out to prove its worth.
At the recently opened Paris Motor Show 2026, QCraft showcased local road test results for its urban NOA. Almost simultaneously, XPENG announced it is opening its smart vehicle technology solutions to the world, pushing R&D capabilities originally meant for its own models into the external market.
Image Source: XPENG
These two events may seem distinct, yet they point to a single shift: China's auto industry isn't just exporting vehicles anymore; it's starting to supply technology to global automakers.
This shift didn't start in Paris. Over the past few months, Chinese autonomous driving firms like Momenta, Shenzhen Zhuoyu Technology Co., Ltd. (ZYT), and Horizon Robotics have been accelerating their entry into the international auto industry through model collaborations, road verification, and technology licensing.
Yet the global automotive supply chain is hardly short on mature technology providers. It remains an open question whether the dominance Chinese firms have built at home can translate into long-term orders from multinational automakers.
That is the question our "Chinese Solutions for Global Markets" series continues to explore.
From European Road Tests to Global Model Development
As Chinese autonomous driving technology goes global, it is evolving from simply following domestic brands abroad to integrating into the R&D systems of foreign automakers.
The partnership between ZYT and Volkswagen offers a case in point.
The recently disclosed ID. Buzz road test in Europe covered 24.9 kilometers over 23 minutes, targeting five typical traffic scenarios: stop lines, temporary construction lanes, roundabouts, right-of-way situations, and speed limits.
This marks the first public demonstration of real-world road testing since the two companies announced their collaboration on next-generation autonomous driving concept verification.
In May, ZYT established its European headquarters in Braunschweig, Germany, and announced the partnership with Volkswagen. Since then, the company has expanded its European footprint by collaborating with commercial vehicle manufacturers like Sinotruk and FAW Jiefang.
From serving Chinese commercial vehicle exporters to participating in the technical verification of European automakers, ZYT is working to build diverse relationships with overseas clients.
Another form of collaboration has already extended into the definition of global models.
Under a global strategic partnership signed by Momenta and Dongfeng Stellantis Automotive Technology (Wuhan) Co., Ltd. (DSAT), four initial new energy models from DSAT are set to roll out starting in 2027, targeting European and other global markets.
Image Source: Momenta
DSAT, established just two months ago, serves as a key operational platform for the transformation of Dongfeng Peugeot Citroen Automobile Company Ltd. (DPCA). The two entities operate under an integrated model with divided functions: DSAT handles R&D, sales, service, and channels, while DPCA manages manufacturing, supply chain coordination, and quality assurance.
Leveraging the industrial resources of Dongfeng Motor and Stellantis, DSAT aims to push new-energy smart vehicles developed in China onto the global stage.
In this setup, Momenta isn't merely supplying an assisted driving system. Per the plan, the company will also participate in defining functional scenarios, developing product experiences, and tuning chassis dynamics—integrating world models with vehicle control capabilities.
Unlike traditional parts supply, this collaboration allows technology suppliers to participate in vehicle development at an earlier stage.
Automakers, too, are beginning to offer their technology to external parties.
XPENG' s newly released XPENG AI ENGINEERING bundles vehicle platforms, modular technology, and tool chains into a commercial solution for global partners. Beyond smart driving and chips, it encompasses capabilities in vehicle R&D, production, and after-sales service.
Technology once reserved for internal models is now being folded into external partnerships. Yet, the scale of specific overseas projects and their commercial returns remain to be seen.
For its part, QCraft' s demonstration in Paris highlights the progress of migrating domestic mass-production technology to the European market.
Image Source: QCraft
According to company disclosures, QCraft QPilot assisted driving solution has been fitted in over 1 million vehicles to date, through partnerships with more than 10 leading domestic automakers across 40-plus models. The European test setup utilizes 11 cameras and one LiDAR, supporting both highway and urban NOA functions.
From road testing and technical validation to joint development of global models, Chinese autonomous driving firms are gaining more opportunities to participate in the international automotive sector.
However, these partnerships are at different stages, and their future commercial scale has yet to be determined.
China's High-End Autonomous Tech: Overseas Buyers Are in No Rush
The pace at which autonomous driving has penetrated the Chinese market is already clear in the data.
In July, Guo Shougang, director of the Ministry of Industry and Information Technology's Department of Equipment Industry, noted that since 2026, the penetration rate of L2 combined driving assistance in passenger vehicles has hit 70.5%, while navigation on autopilot (NOA) features reached 34.2%. The first batch of L3 conditionally autonomous models have also begun operating in designated zones.
In China, driver assistance is expanding from basic functions to more complex piloting scenarios. Yet overseas markets may not follow the same rhythm.
Momenta CEO Cao Xudong suggests that the inflection point for urban NOA in China likely arrived in 2024, while the global market may not hit that mark until 2028.
Horizon Robotics founder and CEO Yu Kai, meanwhile, is more focused on immediate demand.
"Going global is certainly a massive long-term opportunity for autonomous driving," Yu Kai observes. "But in the short term, overseas consumers are primarily looking for ADAS. There isn't a demand for high-level autonomous driving just yet."
Compared to Cao's optimism, Yu estimates that it could take roughly five years for overseas markets to develop the widespread appetite for urban NOA that China sees today.
Their forecasts differ, but they underscore the same reality: The pace of technological advancement in China doesn't necessarily match the procurement cycles of foreign automakers.
Fierce competition at home has compelled automakers to add smart features, giving suppliers a wealth of R&D demand and mass-production opportunities. Abroad, however, the application of driver assistance is shaped by consumer willingness to pay, product costs, regulatory environments, and model development cycles.
Horizon Robotics' overseas order book reflects this divergence.
Yu Kai revealed that as of September, Horizon has secured over 20 million orders for basic driver assistance solutions overseas, with deliveries slated to roll out over the next three to five years. These involve overseas models from Chinese automakers as well as projects from global manufacturers.
Compared to urban NOA, demand for basic driver assistance is more mature, and the path to large-scale delivery is clearer.
This isn't to say high-level autonomous driving has no chance—rather, it requires different strategies for current orders versus future technology positioning.
For Chinese firms, the revenue potential of overseas markets may depend less on the complexity of their technology and more on what customers actually need.
"Opening" a New Market in Months: What Comes Next?
Diverse road conditions abroad were once a major barrier for the export of autonomous driving technology. But with the rise of large model technology, the efficiency of localization is improving.
Previously, Momenta told Gasgoo in an interview that leveraging the generalization capabilities of large models and support from local partners, its solutions can complete initial technical adaptation for a new country or region in about two to three months.
The industry refers to entering a new market with autonomous driving as "opening a country." Taking the Middle East as an example, Momenta notes that models can learn local traffic rules and road conditions within months, making the necessary adjustments.
ZYT's latest European test results demonstrate a different technical approach.
The test vehicles feature an assisted driving solution based on a native multimodal foundation model (MFM). ZYT also employs the real-time controllable world model, ZYT-World, for simulation training and evaluation, followed by real-world road tests to verify results—aiming to boost adaptation efficiency for Europe's complex traffic scenarios.
Both companies are striving to improve the efficiency of technology transfer. Yet for global model development, algorithm adaptation is just one step in the process.
Lv Haitao, general manager of DSAT, points out that much of the past automotive development work centered on the Chinese market. Now, facing Europe, the Americas, Africa, and other regions simultaneously, we must consider regulatory certification, cost requirements, and user needs separately.
"People might use the same words, but the reality is entirely different," Lv said regarding global model development.
This complexity also extends to vehicle integration.
Autonomous driving algorithms must align with sensing hardware, electronic and electrical architectures, steering and braking systems, and software update mechanisms. The Momenta-DSAT partnership, for instance, involves wheel-level precision braking control to enhance stability on low-friction surfaces and improve parking performance.
Commercial vehicles, meanwhile, impose a different set of requirements.
The heavy-truck driver assistance system ZYT showcased in Hanover was developed to address factors like large vehicle dimensions, long braking distances, and significant load variations. It utilizes over 1,000 hours of real-world driving data for model fine-tuning.
Even if the underlying model can be reused across different vehicle types, the dynamic characteristics and safety boundaries of each specific vehicle require recalibration.
Furthermore, overseas mass-production projects must satisfy requirements for functional safety, cybersecurity, data compliance, vehicle certification, and OEM quality systems.
Algorithms may be adapted in a matter of months, but that doesn't exempt suppliers from the rigorous audits by multinational automakers regarding development processes, product liability, and long-term supply capacity.
This explains why Chinese autonomous driving firms are racing to establish local European teams. ZYT has built out an R&D and road testing framework in Braunschweig, Germany, while QCraft has set up a European headquarters in Munich to drive R&D, certification, and business expansion.
For global automakers, a supplier's value lies not just in technical performance, but in the ability to shoulder long-term responsibilities for development, quality assurance, and after-sales support.
Global Supply Chains Aren't Just About Swapping Suppliers
Even with technical prowess and delivery capabilities, Chinese companies must contend with the entrenched supply chain structure of the global automotive industry.
Volkswagen has a long-standing partnership with Mobileye and is jointly developing an L4 autonomous vehicle based on the ID. Buzz. While distinct from the driver assistance concept verification involving ZYT, it highlights that Volkswagen already has an established technical collaboration network.
The same holds true for BMW and Qualcomm. Their jointly developed Snapdragon Ride Pilot is already featured in the new BMW iX3, and the two companies deepened their chip partnership in July to cover vehicle planning for the next decade.
As Chinese suppliers vie for overseas clients, their competitors may already be deeply embedded in the automakers' platform development. Beyond technology and cost, existing software architectures, development tools, quality systems, and established relationships all influence procurement decisions.
However, entering the global supply chain doesn't necessarily mean displacing incumbent suppliers.
Horizon Robotics' business model offers an alternative path.
Yu Kai revealed that roughly 95% of Horizon's overall shipments are delivered through partners, a network that includes international Tier 1 suppliers like Bosch and Denso. Beyond supplying chips and software directly, Horizon participates in automaker R&D through IP licensing.
"Globally, in the two major automotive markets of Europe and Japan, we only have partners, not competitors," Yu says of Horizon's overseas strategy.
This statement reflects its platform positioning, though it doesn't mean Horizon avoids competition with tech suppliers like Qualcomm and NVIDIA. By collaborating with international Tier 1s, Chinese computing platforms can penetrate existing supply chains without a single firm shouldering the entire burden of system development and delivery.
Previously, Gasgoo pressed Yu on this point: securing overseas orders is one thing; truly integrating into a global automaker's R&D and supply chain is another. What defines the completion of globalization for Chinese autonomous driving?
Yu noted that Horizon has established partnerships with global automakers like Volkswagen and Toyota, fostering mutual learning.
"This is no longer a one-way export process; we have become a vital part of the global ecosystem."
From vehicle exports driving supplier expansion to Chinese technology participating in global model definition, system R&D, and industry specialization, the internationalization of Chinese autonomous driving is entering a new phase.
Yet turning these partnerships into sustainable business will require long-term validation of production scale, quality performance, and commercial returns.
Back on the streets of Paris, QCraft's test vehicle is just a microcosm of China's autonomous drive going global. European road tests, partnerships with multinational automakers, and overseas orders are opening more market entry points for Chinese technology.
The real test will come after the first batch of projects: can Chinese autonomous driving secure a place in the next-generation product planning of global automakers?
