Great Wall Motor invests 100 million US dollars in DeepRoute.ai to boost the mass production of high-level intelligent driving.

The urban intelligent driving function is gradually gaining a foothold in users' minds, and traditional brands can no longer avoid this mass production pursuit battle.

36Kr has learned that Great Wall Motor led the Series C financing round of the autonomous driving company DeepRoute, investing 100 million US dollars to become a shareholder. As early as March this year, DeepRoute became the intelligent driving supplier of Great Wall Motor, helping the Wey Lank Mountain model to achieve the mass production and launch of the map-free urban NOA (Navigation on Autopilot) function.

According to 36Kr's exclusive understanding, it is expected that more than 10 high-level intelligent driving models equipped with DeepRoute will be launched in the market next year. After the 100 million US dollars investment, DeepRoute is expected to accelerate the expansion of the mass production scale;

At the same time, DeepRoute is also developing a VLA model (Vision-Language-Action Model) based on Nvidia's Thor chip, and the system will be launched next year.

DeepRoute was established in 2019 and has layouts in both passenger vehicle and light truck commercial vehicle autonomous driving. Currently, DeepRoute has reached mass production cooperation with 4 automakers. Among them, the Wey Lank Mountain of Great Wall Motor and the smart #5 are both equipped with the map-free urban NOA (Navigation on Autopilot) technical solution from DeepRoute.

Intelligent driving is a key project for Great Wall Motor's transformation.

Great Wall is one of the traditional manufacturers in China that has earlier laid out an independent research and development system for intelligent driving. According to Wei Jianjun, the independent research and development history of Great Wall's intelligent driving has been as long as ten years. However, in the industry's mainstream high-level intelligent driving solutions, Great Wall's progress has not met expectations.

The Road of Great Wall's Intelligent Driving: From Independent Research and Development to Outsourcing

In 2019, Great Wall Motor separated its forward-looking department of intelligent driving and established the independent subsidiary Haomo Zhixing, which focuses on intelligent driving research and development. In the following year, Great Wall's "331" strategy was announced, stating that all vehicle models of its 5 sub-brands will be equipped with the intelligent driving products of Haomo Zhixing.

Great Wall Motor is not only the largest shareholder of Haomo Zhixing but also the primary customer.

During its four years of establishment, Haomo Zhixing has successfully mass-produced two sets of intelligent driving products. Among them, the 3,000-yuan-level HP170 can achieve high-speed map-free NOA (Navigation on Autopilot); the HP370, launched in April 2024, can achieve memory driving and parking functions, with a price only at the 5,000-yuan level.

More than 20 vehicle models under Great Wall, such as the Wey Mocha, Tank 300 Urban Edition, and Haval Shenshou, use these two sets of solutions.

Another product, the HP570, positioned at the 8,000-yuan level, will support the map-free urban NOA function. The HP570 uses the Qualcomm 8650 chip and will be first launched on the models of Great Wall Motor. An informed source told 36Kr that both Great Wall and Qualcomm hope that Haomo Zhixing's HP570 can become their mass production benchmark in the field of map-free urban NOA.

However, the progress of Haomo Zhixing has not met expectations, and the project progress of HP570 has not enabled Great Wall to enter the first echelon of map-free urban NOA. Since December 2023, new force enterprises such as Huawei, Xiaopeng, and Li Auto have successively launched the urban NOA function that does not rely on high-precision maps, and currently, the competition focus of the first echelon players has shifted to the mass production and launch of the "end-to-end large model" route.

In Great Wall Motor's plan, the company aims to achieve the goal of accounting for 80% of new energy vehicle sales by 2025. In order to enhance the overall sales of new energy vehicles, Great Wall Motor must quickly catch up with the mass production progress of intelligent driving.

Through the capital path and introducing leading suppliers is the fastest way.

In March this year, Great Wall Motor introduced DeepRoute as its second intelligent driving supplier to provide an end-to-end intelligent driving solution for Great Wall, and the Wey Lank Mountain model is the first collaboration. In 2024, Great Wall will also launch three models equipped with DeepRoute's technical solution.

Great Wall Motor's attitude towards intelligent driving has become more open. In addition to DeepRoute, Great Wall has enabled Moment as the intelligent driving supplier for the Mecha Dragon model and also collaborated with Zhuoyu Technology under DJI.

From independent research and development to outsourcing, Great Wall Motor's pace is obviously accelerating.

Why DeepRoute?

Currently, intelligent driving is increasingly becoming an important decision-making factor for users when purchasing a vehicle. Starting from scratch to independently research and develop intelligent driving takes at least 1-2 years, which cannot win more opportunities for automakers. By adopting the supplier's solution, automakers can strive for the mass production and launch of intelligent driving in less time to win business chips.

However, in the field of high-level intelligent driving, there are currently few intelligent driving suppliers with the mass production capacity of urban NOA. Huawei, Momenta, and DeepRoute are among the few players that have made breakthroughs in the mass production link.

With various business models, Huawei has enlisted automaker brands such as AITO, Zhijie, Xiangjie, Changan Avita, Dongfeng VOYAH, and BYD Fangchengbao.

Momenta has a representative work after in-depth cooperation with SAIC IM L7. Currently, its high-level intelligent driving solution has also expanded to brands such as BYD Denza and GAC AION Hyper.

DeepRoute's understanding of high-level intelligent driving and the end-to-end large model has gained the favor of Great Wall. DeepRoute claims that it has been laying out on this technical route for nearly two years. In August last year, its end-to-end solution started road testing, and this year, it officially launched the high-level self-driving platform DeepRoute IO based on the end-to-end model.

Currently, its end-to-end solution has been implemented on this year's new Wey Lank Mountain; the smart #5, which started delivery in November, also carries DeepRoute's high-level solution.

According to a financing document obtained by 36Kr, DeepRoute has already cooperated with four automakers. DeepRoute disclosed that it is expected that more than 10 high-level intelligent driving models that can be mass-produced in 2025. At the same time, DeepRoute can also promote the algorithm iteration of Robotaxi and unmanned light trucks based on the massive data generated by the pre-installed mass production vehicles.

It is understood that the next-generation end-to-end solution of DeepRoute will also use Nvidia's latest product, the Thor chip, becoming one of the first companies in the industry to use the Thor chip.

The industry consensus is that the "end-to-end" technical route has become the latest direction in the industry. In this natural data and computing power furnace of end-to-end, automakers that take the lead in independent research and development have an obvious first-mover advantage. Players such as Huawei, Xiaopeng, and Li Auto hold massive data and abundant computing power resources, and their end-to-end closed loop has been initially established.

But for intelligent driving companies, when they are inferior to top automakers in data, computing power, and engineering experience, betting on end-to-end intelligent driving is tantamount to a risky gamble.

This also prompts DeepRoute to raise more resources and cash to support and strive for a larger mass production scale, so as not to fall behind in the middle of the end-to-end evolution. And Great Wall can also catch up with the mainstream rhythm of end-to-end intelligent driving in a faster way.