In 1984, China's inaugural vehicle manufacturing joint venture, Beijing Jeep, was formed with American Motors Corporation (AMC) — Jeep's forerunner — as its foreign partner.

Over four decades later, Jeep's new models are poised to feature China's intelligent driving system, aimed at global distribution.

On September 24th, DPCA and Momenta signed a global strategic cooperation agreement. The plan is straightforward: the two parties will collaborate to develop an advanced intelligent driving assistance system for new Peugeot and Jeep models, targeting the Chinese, European, and other key markets.

Historically, when multinational automakers adopted China's intelligent driving technology, it was typically confined to models sold within China. This time, the new models are designed for Chinese, European, and other global markets. According to Lv Haitao, General Manager of DPCA, all markets except the United States are, in principle, open.

Corporate roles are evolving too. In this partnership, Momenta is not merely a supplier providing a standardized package of functions to automakers; instead, both parties are working together to define the product.

Momenta CEO Cao Xudong stated in a media interview, "This cooperation holds dual significance in terms of progress. First, it marks the transition from conventional to intelligent for the joint venture brands Peugeot and Jeep. Second, the concept of 'in China, for the world' was previously unimaginable; it was more about 'in China, for China'."

Momenta CEO Cao Xudong

Over the past thirty years, the underlying logic of the joint venture model has involved exchanging market access for technology. Overseas entities provided technology, platforms, and brands, while China offered market scale. Now, China is not just a recipient of new technologies but has become a hub for the research and development of global vehicle intelligence.

From this perspective, the DPCA and Momenta collaboration can be seen as a test of whether China's intelligent driving solutions can expand from local to overseas markets.

One Intelligent Driving System, Customized for Different Brand Identities

In traditional auxiliary driving collaborations, suppliers typically provide sensors, domain controllers, software, or functional packages, while automakers handle vehicle integration, testing, and market launch.

This model is efficient but tends to treat intelligence as a standardized component.

In the DPCA and Momenta collaboration, Momenta provides driving decision-making capabilities, enhancing vehicle intelligence, while DPCA is responsible for the chassis and driving experience, translating the design intent of intelligence into vehicle actions.

The intelligent driving experience of the new models is jointly defined by both parties, encompassing cross-scenario intelligence from urban to off-road driving, as well as integrating AI decision-making with chassis control for precise system adaptation.

Why doesn't DPCA undertake this alone? Lv Haitao has assessed the costs:

"Relying solely on an automaker's capabilities to support an intelligent driving assistance system is, in my judgment, insufficient in terms of scale, capability, and investment. It's challenging for us automakers to keep up with this level of demand."

This is the rationale behind the joint development model. At the same time, joint development can also leverage the mechanical expertise that Peugeot and Jeep have accumulated over decades.

Peugeot's core strength lies in its driving engagement, characterized by its ability to translate professional racing-level chassis tuning into mass-produced vehicles. Jeep excels in four-wheel drive and off-road capabilities, making all-terrain performance a brand tenet and delivering a sense of limitless freedom.

For such brands, excellent intelligence is a bonus, while poor intelligence means algorithms replace the soul. Many long-established automakers worry that after installing intelligent driving systems, their vehicles will no longer feel authentic.

However, Momenta's collaborative approach involves calibrating algorithms with traditional automotive engineering. It aims not only to preserve the brand characteristics of Peugeot and Jeep but also to continue their unique French driving control and American off-road personalities in the intelligent era.

Such complementary collaborations are not unique. Previously, BMW Group China reached a cooperation agreement with Momenta, also emphasizing the integration of German driving philosophy with intelligence.

This is also the breakthrough of this collaboration: intelligence should not make cars identical but should make each brand more distinct.

How a Single World Model Serves Global Markets

Lv Haitao vividly recalls Momenta's technology.

In January of this year, Cao Xudong brought an intelligent driving test vehicle to DPCA, and the two spent a full hour and a half driving on various roads in the development zone, with the self-driving car unaffected by their conversation.

DPCA General Manager Lv Haitao

To achieve this, Momenta relies on its R7 World Model.

Traditional intelligent driving solutions teach vehicles to remember more driving scenarios so they know how to handle them when encountered again. In contrast, the R7 model is self-learning, enabling vehicles to understand why the physical world moves as it does and then think about how they should act.

The technical path of this model consists of three layers: pre-training with real driving data, closed-loop simulation of the world model, and reinforcement learning.

Simply put, in the first two layers, Momenta feeds the model with massive amounts of real driving data to help it understand the true laws of the physical world. It then uses the pre-trained world model as a virtual sandbox to let the model simulate how its actions will produce changes.

The third layer builds a highly realistic virtual training ground based on the first two layers, allowing the system to learn from mistakes and optimize. Ultimately, it finds the optimal solution in extreme scenarios.

However, this does not mean that a single fixed algorithm can be directly deployed in all global markets.

Regulations and data pose significant challenges. The EU's GDPR strictly restricts road data collection and cross-border transmission, meaning the data flywheel that works in China may not spin in Europe.

Additionally, differences in roundabout yield rules in Europe and varying driving habits across regions require intelligent driving companies to rebuild not just parameters but also decision-making logic.

This means that taking intelligent driving overseas is never about software replication but a long-term project involving a unified base, local data, vehicle adaptation, and compliance verification.

Momenta began preparing early. In 2021, it established overseas offices in Germany and other locations, later expanding to other regions. Overseas data reserves and data center infrastructure construction began at that time and had become relatively mature by 2023-2024.

Cao Xudong stated, "Even compared to overseas self-driving companies, we are at least one step ahead in terms of compliance with overseas regulations."

This early investment is now paying off. In July 2026, Momenta obtained a nationwide L4 testing permit from the German Federal Motor Transport Authority (KBA). This is the first time a Chinese company has received such a national-level qualification without needing to apply city by city.

One detail is that DPCA was initially concerned about whether Momenta's solution could handle the complex road conditions near the Arc de Triomphe in Paris. After hearing this, Cao Xudong immediately showed a video of a vehicle passing through the Arc de Triomphe.

Behind this is Momenta's overseas accumulation and the technological dividends brought by its data flywheel. Public data shows that Momenta has moved beyond single-project validation. With over 1.2 million mass-produced vehicles delivered, a cumulative 17 billion kilometers of real-world driving, and 100 million high-value training data segments, it means that even larger volumes of mass-production data are supporting model iteration.

Today, Momenta's R7 World Model is providing a new path for China's intelligent driving technology to go global, using a unified world model base platform to integrate multi-market experience and leveraging mass-production closed loops to drive continuous technological upgrades.

Chinese Technology Feeds Back to the Global Market

The global automotive industry's division of labor is undergoing profound changes, shifting from foreign companies entering China to establish joint ventures to Chinese technology going global.

In the past, the narrative was that China provided the market, while foreign companies provided technology and platforms, with global vehicle models being introduced for production in China. Brands like Dongfeng Nissan and SAIC-GM have all adopted Momenta's solutions, but primarily to serve Chinese consumers.

The DPCA deal reverses this dynamic. Lv Haitao mentioned, "This time, our shareholders have given us significant autonomy, so we propose a new era of joint ventures — an era of autonomous joint ventures."

In terms of technological pathways, Momenta uses foresight to assist international brands in pre-emptively deploying global mass production of advanced intelligent driving systems.

For DPCA, this collaboration clarifies the technological pathway for developing intelligent driving systems for new Peugeot and Jeep models, eliminating the need to build a self-driving system from scratch and allowing direct access to cutting-edge physical AI technology.

In terms of global expansion, relying on Stellantis' global marketing channel network, China's advanced driving assistance solutions are expected to achieve large-scale global coverage, with Momenta's solutions directly reaching overseas markets through new mass-produced models from Peugeot and Jeep.

Lv Haitao stated that this expansion speed directly benefits from the parent company: "They already sell cars globally. It's not just about sales resources; their work systems for local regulatory certifications in various regional markets are already well-established."

Cao Xudong predicts that the inflection point for China's urban NOA market will occur around 2024, while the global market may see a similar shift around 2028.

Based on this prediction, Momenta aims to initiate vehicle development and overseas deployment preparations ahead of time, at least half a step earlier than the inflection point.

Its confidence comes from multi-dimensional reserves in mass production and data. Momenta currently holds the largest market share in China's third-party urban NOA segment. According to CIC data, from March 2025 to February 2026, Momenta accounted for approximately 65% of the market share among third-party urban NOA suppliers.

Momenta also has a strong hand. With 1.2 million vehicles delivered, 17 billion kilometers of real-world driving, and 100 million training data segments accumulated; international automakers like Mercedes-Benz, BMW, and Volkswagen have established collaborations with Momenta through joint development and model designation; backed by the intelligence Moore's Law, hardware costs halve every two years, while experience improves tenfold every two years.

As global user acceptance of advanced driving assistance systems increases, this first-mover advantage has the potential to be replicated overseas.

The collaboration models are planned for gradual launch starting in 2027, just ahead of the global inflection point window in 2028. For overseas consumers, Momenta offers a trustworthy and verifiable solution that has been validated through mass production on a scale of millions of units.

For the global automotive industry, the division of labor is being rewritten. In the past, overseas entities defined the standards, and China adapted. Now, Chinese brands have gained the right to define standards, serving global consumers, marking a new cycle in history.