搜尋歷史
熱門搜尋
經銷商登入 / 註冊
In the past two years, the choice of intelligent assisted driving solutions by automotive companies has triggered a massive discussion in the industry regarding the "Soul Theory"—but this has not stopped most automotive companies from seizing market opportunities. They continue to choose mainstream suppliers' intelligent driving systems and smart cockpits to improve their products' competitiveness. Obviously, establishing an intelligent driving system requires evolution from "consciousness" to "pragmatism," and it cannot be achieved overnight, as this involves R&D investments in the range of tens of billions. Now, a new Central SOE has finally taken a substantive step, and that is Changan Automobile.
On September 4th at the 6th Changan Tech Ecosystem Conference, the release of Changan Tianshu Leading technology and the pre-sale launch of Changan Qiyuan Q06 were officially held, which can be considered a closed-loop response to the "Soul Theory" of that year. In the subsequent media communication session, Changan Automobile Executive Vice President Yang Dayong directly pointed out the core of Changan's intelligent driving self-research: "From the beginning of creating the Changan brand and Qiyuan brand until now, what we were waiting for was Tianshu Leading. Changan Qiyuan models will be equipped with, and only equipped with, Changan's self-developed Tianshu Leading intelligent driving assistance system in the future. The core reason comes down to two words: Brand."
Objectively speaking, in this era where almost every car company talks about "end-to-end," there are not many companies that have truly implemented mass production with a fully self-developed one-piece end-to-end approach. However, Changan is an exception among them. With cumulative sales breaking through 30 million units, persisting in intelligent R&D for over a decade, and investing over 60 billion into the intelligent sector, Changan's firmly chosen path of full-stack self-research still holds relatively high value today.
Of course, if this path is simply labeled as "computing power competition" and "scenario coverage," it is actually looking at Changan's decision too simply. Firmly choosing such a capital-intensive route and releasing a product with superior intelligent driving levels without setting a high threshold, the strategic significance behind Changan's intelligent driving self-research is worth deep discussion.
Changan Intelligent Driving Self-Research Understands the "Coase Boundary"
Even though intelligent driving is a topic that mainstream car companies must face, the matter of "whether to self-research" is understood differently by everyone. Facing fierce market competition, thinking about your own "Coase Boundary" in advance is perhaps the more reasonable answer to this question. The Coase Theory holds that when market transaction costs are higher than internal organization costs, enterprises choose to "do it themselves" rather than "buy from others." In other words, how much an enterprise should do depends on the cost comparison between building in-house and outsourcing.
Theoretically, the current Chinese intelligent driving market has a fairly mature supplier ecosystem. Any of Huawei, Momenta, or Horizon could allow Changan to take a shortcut in intelligent driving; Changan has the condition for "plug and play" in this field. Transferring R&D costs to the supply chain and focusing on vehicle manufacturing and brand marketing is also a choice for many car companies.
Surprisingly, Changan did not do this. So what was its choice? Since establishing an intelligent R&D team in 2009, Changan has persisted on the self-research route, building an intelligent R&D team of over 7,500 people and constructing the only National Key Laboratory in China focusing on intelligent vehicle safety technology. It can be seen that Changan's investment in this area is huge.
"Why must we insist on self-research? Without self-research capability, it is difficult for the brand to clearly explain to consumers: Why should users choose us? What differentiated experience can we provide? Consumers are not just buying a car, but an affirmation of the brand, which is also an important part of our brand's new quality productive forces." Yang Dayong stated.
Then, why take the detour instead of the shortcut? In our exclusive interview, Changan Automobile Vice President He Gang provided an explanation. He mainly mentioned four points: First, only self-research allows for deep understanding of user needs and polishing products with real data; Second, only self-research possesses complete system analysis capabilities. When vehicle anomalies occur, problems can be located at the whole vehicle level. If only doing the intelligent driving single module, the whole picture of the vehicle cannot be seen; Third, only self-research can achieve rapid iterative optimization. After problems are found, data can be quickly collected, models trained, and simulations verified; Fourth, only self-research can back up safety and responsibility. As an OEM, it must be responsible for the final result of the whole vehicle.
From these perspectives, Changan's decision logic is very clear. When intelligent driving technology becomes a "core differentiated element" of automotive products, market transaction costs will exceed internal organization costs. Outsourcing intelligent driving solutions may look like a "shortcut" on the surface, but hidden costs are very high. Apart from negotiation costs, coordination costs, and iteration costs, the core thing belongs to others, making it difficult to form a true technical barrier. For a car company of Changan's scale, "borrowing the path" is not a long-term plan.
Besides the underlying considerations at the strategic level, the long-term value of the self-research route is more reflected in the core advantages of the technical architecture. Changan Automobile Chief Intelligent Driving Technology Officer Tao Ji broke down Tianshu Leading's "one-piece end-to-end" into three layers: First is the World Model. Driving is a dynamic game task; the vehicle needs to truly "read" the intentions of other traffic participants and play games with them; Second is the Body Model, incorporating vehicle dynamics and chassis capabilities into the large model to understand the "boundaries of muscle capability"; Third is the Multimodal Large Model, allowing the system to understand natural language instructions like "follow that grey sedan in front" and accurately map them to specific vehicles in 3D physical space. The three eventually share a backbone of a basic large model in the cloud. This is also the most essential difference between Tianshu Leading and "outsourced solutions."
Changan chose to self-develop Tianshu Leading. This system forms a deep coupled system-level synergy with the SDA Tianshu Central Ring Network Architecture, Tianshu Chassis, and Tianshu Cockpit 2.0. This synergy cannot be achieved by any outsourced solution. Yang Dayong stated at the communication meeting: "According to Changan's planning, whether it is Changan Automobile or Changan Qiyuan Brand, Changan brand models will be equipped with, and only equipped with, Changan's self-developed Tianshu Leading intelligent driving assistance system. So far, we have no intention to introduce other suppliers to provide intelligent driving assistance overall solutions for the Changan Brand. The core reason comes down to two words: Brand." Changan wants to clarify "why choose us" to consumers through this way and further shape brand recognition. From a long-term perspective, this is a very visionary move to enhance brand competitiveness.
Changan's ambition goes beyond self-research and self-use. He Gang revealed: "When Tianshu Leading was originally planned, it was to build a global intelligent platform, hoping to provide it to all car factories and industries related to intelligent robots, such as robots, flying cars, etc. Currently, besides our product brands, external car factories are also communicating with us. Some strategic agreements will be released recently, and next year, some models in vehicles outside the Changan family will be equipped." This means Changan's role positioning in the intelligent driving field is extending from "Client" to "Solution Provider."
Behind Qiyuan Q06 Pricing is a Downward Marginal Cost Curve
At the 6th Changan Tech Ecosystem Conference, Qiyuan Q06 started pre-sales. The pre-sale price starting from 147,900 RMB was quite unexpected, especially since this is a vehicle fully equipped with LiDAR and carrying the Tianshu Leading intelligent driving system. If adding the 800V Silicon Carbide High Voltage Platform, 6C Fast Charging, front double wishbone rear multi-link full aluminum suspension, etc., it is clearly a pricing exceeding industry norms.
Now the industry is heavily competitive, and breaking existing price anchors while improving value experience is not easy. Now Changan has done this with this new car. Why? After all, its marginal cost curve is moving downward. The cost structure of outsourced solutions is "linear," making the upfront investment look acceptable. However, in reality, for every car sold, the car company needs to pay a fixed licensing fee or hardware premium to Tier 1.
To put it bluntly, no matter how much car companies want to reduce costs, there is a supplier's profit floor controlling it. Self-research solutions are different. The cost structure of self-research solutions is "steep then flat," meaning huge upfront investment, but the marginal cost keeps decreasing. For example, Changan's upfront R&D investment is a fixed cost. Once mass production is completed, the "intelligent driving marginal cost" for every additional car is only chip, sensor, and other hardware material costs.
Beyond self-research, Changan has further confidence in cost reduction. Hesai's official website mentioned that it has obtained exclusive mass production projects for multiple Changan brands, with planned orders exceeding 1.5 million units. This scale of centralized procurement means Changan has more bargaining space on the hardware side. Therefore, behind Yang Dayong saying "this car must make money," is a problem calculated in advance. On one hand, self-research amortizes fixed costs; on the other hand, centralized procurement suppresses hardware costs. Under dual-wheel drive, Changan dares to break norms with Qiyuan Q06 and re-calibrate the value of this level of vehicle. Qiyuan Brand's actual sales also prove the correctness of Changan's strategy in recent years.
"Qiyuan's global sales broke 45,000 in August. After Q06 is out, I believe the number will increase further. Moreover, this number was achieved under the difficult situation of guaranteeing Q05 supply now, with tens of thousands of Q05 orders still not delivered. Plus Q06, I personally think Qiyuan's average monthly sales might have a chance to reach around 60,000 units, and the total sales this year should be around 400,000 units." For Changan Qiyuan's current achievements and next goals, Yang Dayong believes there is more confidence and optimism with the support of Tianshu Leading.
In the interview, Changan Qiyuan General Manager Li Rui revealed that Q06 pre-sales reached 11,000 units in two hours last night. Among the two major configuration gradients, one is Tianshu Leading Pro, the other is Tianshu Leading Ultra. The percentage of customers choosing Tianshu Leading Ultra reached 53%. Most customers chose the 700 km range, and they have a higher focus on intelligent driving assistance. Especially Changan Automobile's inaugural dual-motor rear-drive data, the users who chose it account for about 33%. This means the willingness of initial users to pay for intelligent driving is not low, and it also shows that although car companies have huge investment in self-research, the outbreak of market demand will allow car companies to quickly welcome the return period.
Gaining User Trust, Central SOE Identity is an "Open Card"
To take root in the intelligent driving track, "safety" is the story that must be told well. However, various intelligent driving safety accidents that occurred in the industry in the past did not leave people with a good impression. The trust deficit consumers have in intelligent driving is a problem every car company must face. Changan's advantage lies in its ability to tell the story well and its strength to implement the story. Of course, here is an "open card," which is the Central SOE identity.
In economics, there is a concept called "Credible Commitment." When one party can make an irrevocable commitment, transaction costs can be reduced, and trust efficiency improved. As a Central SOE, Changan has a very high cost to breach commitments. This "dare not mess around" constraint actually makes it easier for people to generate trust. At the launch event, Changan repeatedly emphasized "Assisted Driving is not Autonomous Driving" and "No Exaggeration, No Misleading." Its statement on intelligent driving safety issues carries the characteristics of "Credible Commitment."
This "dare not mess around" constraint is also reflected very specifically in Changan's R&D system. He Gang revealed at the communication meeting that Changan has officially included intelligent driving "generalization testing" into the development test system. 500 engineers cover 45 cities nationwide for road tests. After each version iteration, engineers receive differentiated map tasks for on-site testing. The backend statistics takeover times and other key data in real-time. This mechanism ensures Tianshu Leading will not only stay at "works well in Chongqing," but truly possesses "can work nationwide" generalization capabilities.
In addition, there are a few key points that can indirectly reflect the trustworthiness of Changan's intelligent driving. Changan is the first car company in China to obtain official formal license plates for Level 3 autonomous driving and is also one of the first Central SOEs to obtain Level 4 Robotaxi test licenses. Behind these two licenses are the strict audits of ministries such as the Ministry of Industry and Information Technology and the Ministry of Public Security, carrying the meaning of "Official Certification" in itself. Changan also deeply participated in the formulation of ADS GTR International Regulations—the world's first unified technical regulations for autonomous driving—and led the formulation of multiple national standards such as "Automated Driving Classification of Automobiles." Currently, Changan has deeply participated in formulating over 270 international, national, and industry standards, including over 60 in the intelligent field. It can be seen that Changan's strength in intelligent driving is very hardcore.
Besides, the Qiyuan Brand has not only focused on the domestic market but repeatedly emphasized "Domestic and Overseas Global One Game." In August 2026, Changan Automobile's overseas deliveries reached 92,400 units, up 78.6% year-on-year, achieving year-on-year growth for 7 consecutive months; With the completion of the Brazil Factory production and the 20,000th whole vehicle rolling off the line from the Thailand Rayong Factory, Changan's localized production network continues to deepen, and the overseas market is becoming one of its core growth pillars. He Gang's judgment on overseas intelligent driving implementation is that Highway NCA will spread quickly, but Urban is not that easy. After personally visiting multiple countries, he believes that the UAE road conditions are clear, lane lines are neat, and driving behavior is good, which might be the best overseas market for layout Urban NCA. Senior management is not aggressive, cautious and rhythmic attitude, which is very rare in the industry.
Trial of multi-form consumer markets has already become an important idea for Changan Automobile after core technology support, because for a time-honored car company, the "sense of crisis" always exists.
Yang Dayong analyzed: "There are currently about 120+ automotive brands in China, among which there are about 50-60 with certain sales scale that can enter the rising list. Among these 50-60 brands, the truly active ones, which contribute about 80% of China's auto market sales, are about 25, belonging to 10 to 15 OEMs. According to the '80/20 Rule' deduction, by 2030 or a future point in time, the ones that may finally survive are just these 20+ brands. From over 120 reducing to 20+ means the Chinese market will eliminate about 100 brands in the next few years, inevitably followed by extremely fierce even brutal competition." According to his judgment, before 2030, competition in China's auto market will still be in a white-hot stage. Changan Automobile is confident to be among those 20+ brands and can persist in walking for a long time, and this requires enterprises to build unique competitive advantages in brand power, product power, marketing power, and other aspects.
Conclusion
Tianshu Leading launch, redefined the marginal cost curve of intelligent driving with full-stack self-research and scale effect; Qiyuan Q06, with a pre-sale price starting from 147,900 RMB, reshaped the value anchor of intelligent driving models; Plus the Central SOE identity trust endorsement, management's control of industry rhythm, and constructed differentiated brand cognition, Changan is building a differentiated brand cognition system, every step is walked solidly and clearly.
Industry competition will only become more fierce, and Changan's answer sheet about the "Soul Theory" has been written clearly enough. This is not only a business choice for a car company to gain market recognition through technical self-research, but also a strategic layout for a new Central SOE to carry the banner of core technology independence and controllability in the automotive intelligence wave. From self-research and self-use to open empowerment, from the domestic market to the global stage, Changan's strategic layout ultimately points to the core discourse power of China's automotive industry in the global intelligent competition.
