The Battle on Another Track: Domestic Autonomous Driving Companies Besieging Tesla

In October 2025, Tesla officially rolled out the FSD V14.2 version of its assisted driving system.

Just two months later, XPeng Motors CEO He Xiaopeng personally traveled to North America to experience it, and after a 4-hour test drive of Tesla’s FSD V14.2, he said it brought him close to L4 autonomous driving.

Under the pressure of Tesla’s FSD, domestic intelligent driving companies haven’t been idle, launching surprise attacks on the track.

First, on March 2, XPeng released its second-generation VLA, prompting brokerages like Morgan Stanley and HSBC to give “Buy” ratings.

By March 5, domestic intelligent driving company Momenta quietly filed for an IPO with the Hong Kong Stock Exchange.

Huawei is also expanding its circle, such as the new FAW-Audi A6L released on March 10, which is equipped with Huawei’s intelligent driving system.

On one hand, Tesla’s impressive FSD V14.2; on the other, the relentless efforts of Chinese intelligent driving companies.

All signs indicate that the competition in new energy vehicles has become increasingly clear: intelligent driving is the core, and whoever can lead in this field will define the next generation of cars.

A grand battle to challenge Tesla has quietly begun.

Suppliers and In-House Developers

Building a Complete Intelligent Driving Map

Chinese intelligent driving companies attacking Tesla can be divided into two factions.

One faction consists of suppliers of intelligent driving assistance systems, working closely with automakers to advance system evolution. The other faction comprises automakers developing their own technology, choosing to keep the technology firmly in their own hands and continuously building their barriers.

Among suppliers, Huawei and Momenta are undoubtedly leading the charge.

As of March 3, vehicles equipped with Huawei’s HarmonyOS-based QianKun ADS have avoided 3,540,000 potential collisions. Their intelligent driving system has become a key technical asset.

More importantly, their systems are not only used in Huawei’s HarmonyOS vehicles but are also the preferred choice for many domestic and international automakers.

Companies like Aeva, DeepBlue, Lantu, Audi, and Toyota are among those. You’ll find that not only domestic firms have joined Huawei’s ecosystem; even foreign brands once dominant in fuel vehicles are beginning to discuss cooperation with Huawei.

Momenta is another representative supplier, following a more “Android-like” approach.

It adopts a deep collaboration model with automakers, customizing intelligent driving systems for different models, using data-driven large models to attract cooperation, and continuously feeding back vast amounts of data to improve the models, forming an automated closed loop.

This model has strong irreplaceability. Traditional suppliers typically offer a standard solution for automakers to adopt, like selling standard parts. Momenta, however, brings algorithms and works together with automakers to modify parameters as needed. If an automaker has specific requirements, Momenta will create a dedicated branch.

In this scenario, switching intelligent driving solutions involves high costs in data migration, model adaptation, and team integration. This allows Momenta to quickly capture third-party markets, making it a leading enterprise among intelligent driving suppliers.

On the automaker side, most develop their own technology, with XPeng being the most representative.

Perhaps inspired by the test drive of Tesla’s FSD V14.2 at the end of 2025, or due to accumulated expertise, XPeng officially launched its second-generation VLA on March 2, 2026.

The second-generation VLA differs from the previous one, targeting L4-level autonomous driving, and uses a native multimodal physical world large model. Its processing speed is increased 12-fold, safety takeover mileage is 50 times longer, average takeover distance is 25 times longer, and it achieves capabilities comparable to FSD’s latest features.

Besides Momenta, Huawei, and XPeng, many other companies are working hard in this track. These two factions together form a comprehensive map of China’s intelligent driving landscape.

This map has a distinctive feature: whether supplier or in-house developer, they are increasingly recognized and chosen by foreign brands.

For example, Audi collaborates with Huawei; Toyota partners with Momenta; XPeng’s second-generation VLA and Turing chips have also been selected by Volkswagen.

All signs show that China is ready to compete head-to-head with Tesla in intelligent driving.

Advantages in Competing with Tesla

Many wonder why Chinese intelligent driving has been able to leap from trailing to challenging Tesla within just a few years. What are the advantages in this competition?

In fact, China’s rise in intelligent driving results from the combined effects of policy support, industrial chain maturity, and market dynamics—each rooted in solid accumulation rather than chance.

First, policy guidance is the core support for Chinese intelligent driving.

On April 2, 2025, the National Development and Reform Commission, Ministry of Industry and Information Technology, and four other departments jointly issued the “First Batch of Large-Scale Application Pilot Notice for Vehicle-Grid Interaction,” officially opening access for L3 autonomous vehicles, selecting Shanghai, Beijing, Shenzhen, and seven other cities as pilot zones.

This policy provides room for error and development opportunities for Chinese intelligent driving companies. In contrast, Tesla faces limitations due to cross-border data policies, preventing full utilization of local Chinese data to optimize algorithms—its biggest weakness.

Second, the comprehensive maturity of the industrial chain is a fundamental guarantee.

The development of any industry depends on a mature supply chain. For intelligent driving to be practical, it’s a manufacturing issue: without chips, computing power is meaningless; without sensors, perception is blind. These components cannot be produced by a single company alone; they require a complete industrial ecosystem.

China’s intelligent driving industry chain is already quite mature.

In chips, companies like Horizon’s Journey series, XPeng’s AI Turing chips, and Momenta’s BMC 7X directly compete with NVIDIA’s Orin series.

In sensors, companies like Hesai Technology and Sagitar have made lidar affordable—what once cost over ten thousand yuan and was only installed in luxury cars now costs in the thousands, suitable even for 100,000-yuan new energy vehicles, giving China a significant advantage.

The midstream integration solutions are numerous, with both in-house and supplier approaches contributing significantly—these engineering capabilities are built through countless project experiences.

Finally, mass production capacity is a key strength. China has the world’s most complete new energy vehicle supply chain—from batteries to motors to electronic controls—all domestically produced. This means a complete intelligent driving solution can be quickly matched with compatible models, and driving data can rapidly feed back into large models, forming a closed loop.

In contrast, Tesla lacks this advantage. Despite impressive FSD performance in North America, its data training cannot be directly applied to Chinese roads.

For example, XPeng’s second-generation VLA has been tested extensively in urban villages, alleys, and mixed traffic scenarios, outperforming Tesla’s FSD.

Overall, this self-controlled industrial chain acts as a backbone for Chinese intelligent driving, enabling companies to achieve technological breakthroughs at lower costs and faster speeds.

Finally, precise understanding of user needs is crucial for breaking through.

Tesla’s intelligent driving solutions are standardized global products, with full-feature costs reaching 64,000 yuan. In contrast, most Chinese domestic brands choose to include intelligent driving as standard, and even if charging extra, prices are far below Tesla’s.

Since intelligent driving is still at L2, far from the final stage, users are reluctant to pay for semi-finished products. Autonomous brands’ low-threshold pricing makes them more acceptable. When L4-level autonomous driving arrives, consumers will already have developed a preference for domestic brands’ intelligent driving. This practical pricing reflects a keen insight into user needs.

These three structural drivers give Chinese intelligent driving companies the confidence to challenge Tesla. Their rise is not accidental but the inevitable result of policy, industry chain, and market forces working together.

Epilogue

After all this, when will China surpass Tesla’s FSD?

In reality, the key is not when but whether we have the confidence to compete.

In the past automotive market, Tesla’s FSD was like a banner—whether it was the end-to-end era of FSD V12 or the stunning visual-only approach, Tesla proved its strength.

But now, Chinese intelligent driving companies have shifted from followers to challengers. Tesla is powerful, but no longer invincible.

This equal footing itself is the greatest confidence.

So, returning to the question: when will we surpass FSD?

Perhaps the answer is: the surpassing is already happening.

In every domestically produced new car leaving the factory, in every algorithm iteration, in every signing ceremony with foreign brands, and in every moment a driver activates the intelligent driving feature.

Tesla has indeed ignited the spark, but that spark has already grown into a wildfire.

I believe that the beginning of the intelligent driving chapter was written in Silicon Valley.

But the ending will be written by China’s intelligent driving.