In August 2026, a statement by Yong Jun, Vice President of Avatr, regarding the cooperation model with Huawei sparked heated discussions. Subsequently, Avatr officially issued an urgent clarification: the relevant statements were "misinterpreted," and what Yong Jun meant was that "there is no need for Yinwang, a subsidiary of Huawei, to create a dedicated intelligent driving version exclusively for Avatr," rather than implying that cooperation with Huawei is unnecessary.
This controversy itself has precisely become an excellent window to observe the industry's mindset. From SAIC Motor's "Soul Theory" in 2021 to today's "non-essential" discussion triggered by Avatr, over the past five years, Huawei has evolved from a "trial" option for a few automakers to an almost "standard" presence today. Meanwhile, the oscillation of automakers between "embracing" and "breaking free" has never ceased.
Every step of Huawei's "publicization" is forcing automakers to answer that old question: when the "bodies" become increasingly similar, where exactly does the "soul" reside?
One "Non-Essential", Two "Survival Instincts" The cooperation between Avatr and Huawei covers multiple levels, including capital, technology, and products, making the discussion triggered by this controversy more exemplary. Avatr is the first partner brand under Huawei's HI (Huawei Inside) model, and also the second-largest shareholder of Yinwang. In August 2024, Avatr acquired a 10% stake in Yinwang for RMB 11.5 billion. By October 2025, the entire investment had been paid in full.
Over the six years of cooperation between the two parties, Huawei's mark is visible almost everywhere in Avatr, from brand positioning to product definition.
After the topic of the "Huawei Necessity Theory" quickly sparked discussions, on August 10, Avatr officially issued an urgent clarification regarding the "misinterpretation," stating that what Yong Jun meant was "there is no need for Yinwang to create a dedicated intelligent driving version exclusively," rather than implying that cooperation with Huawei is unnecessary.
Almost simultaneously, Zeng Qinglin, General Manager of the Yijing brand, posted on Weibo to express his stance, stating that Yijing aims to be a "tech geek for Huawei-ecosystem vehicles," emphasizing "the irreplaceability of Huawei in the fields of intelligent driving, smart cockpit, and automation." He also stated that Yijing and Huawei "have had no 'spare tire mentality' since the very first day of full-stack native co-creation," and both parties "bring out their strongest assets, combine them, and deliver them to users"—an "all-in" deep integration.
The juxtaposition of these two postures reflects the collective anxiety of China's automotive industry: when Huawei transforms from a "lifesaver" into "public infrastructure," how exactly should "the necessity of cooperating with Huawei" be defined?
To answer this question, it must be examined within the coordinates of the six-year evolution of Huawei's automotive business.
According to official disclosures from Huawei, as of July 2026, Qiankun ADS (Qiankun Intelligent Driving) has cooperated with over 25 brands and more than 50 car models, with the system's vehicle installation volume exceeding 1.9 million units. Jin Yuzhi, CEO of Huawei's Intelligent Automotive Solution Business Unit, revealed in July 2026 that its dedicated R&D investment for 2026 exceeded RMB 18 billion, the number of equipped models is expected to exceed 80 by the end of the year, and the cumulative installation volume is expected to reach 3 million units. Jin Yuzhi positioned Huawei as "electronic screws," focusing solely on core intelligent components.
However, when this "screw" is screwed into almost all price segments, from RMB 150,000 family sedans to million-RMB luxury cars, the other side of "publicization" is the sharp escalation in the difficulty of differentiation. Huawei seeks scale, while automakers seek uniqueness—this contradiction is surfacing in the current discussion.
"The fact that the 'Huawei content' has become a standard feature itself signifies the arrival of an inflection point in the industry's intelligentization," an analyst from Gasgoo Research Institute pointed out. "When everyone has Qiankun ADS, true differentiation must instead return to the automakers themselves."
The analyst further explained that the so-called "Huawei content" is not a single dimension; there are at least two levels of distinction. First is the depth of integration: some automakers only introduce one or two products, while others achieve full-stack coverage from the smart cockpit to intelligent driving. Second is the technological generation: even within Qiankun ADS, there is a generational gap between ADS4 and the latest ADS5. The analyst explicitly stated, "What Huawei provides is a capability foundation, but how to tune this foundation to bring out different 'personalities' varies in understanding among different automakers."
A deeper tension lies in the cost structure. Taking Seres as an example, the annual sales of the AITO series leaped from less than 10,000 units to over 400,000 units, but it is expected to incur a loss of RMB 1.5 billion–1.8 billion in the first half of 2026. According to industry analysis, Huawei's cooperation with automakers usually adopts a "basic service fee + sales commission" model. The hardware cost for intelligent driving is about RMB 15,000–30,000 per set, with an annual software licensing fee of several thousand RMB and a sales commission rate of 10%–15%. For models priced above RMB 300,000, Huawei may gain a total revenue of RMB 30,000–50,000 per vehicle sold.
Deep binding brings technological leadership and brand endorsement, but it also means compressed profits and diluted brand independence.
However, Huawei's "publicization" will not shift according to the will of a single automaker. From the first AITO M5 in 2021 to the parallel operation of the "Five Realms" under Harmony Intelligent Mobility Alliance (HIMA) and the expansion of the HI model to 25 brands in 2026, Huawei has completed the leap from "exclusive supply" to a "public platform." Its role has also transformed from a selected "supplier" to "infrastructure" that companies are eager to connect to.
For new brands, cooperation with Huawei can be described as timely help—without technological endorsement, it is hard to gain initial trust. However, for enterprises that have already established brand recognition, over-reliance may instead dilute brand distinctiveness.
However, for some automakers, the cost of lacking Huawei's empowerment is equally high. As of August 2026, the cumulative assisted driving mileage of Huawei's Qiankun ADS has exceeded 13.7 billion kilometers. This accumulation of data cannot be replicated by any single automaker in the short term. The prerequisite for "creating differentiated things that others don't have" is sufficient technological confidence—which is precisely the scarce resource today.
Thus, a paradox emerges: the more successful Huawei is, the thinner the "uniqueness" of individual automakers becomes; yet without Huawei's empowerment in intelligentization, most automakers struggle to maintain competitiveness in the intelligentization track. This paradox is the deep reason why the Avatr controversy has resonated widely. Obviously, this issue is no longer just a strategic choice for a single enterprise, but a microcosm of the collective anxiety of the entire industry in the deep-water zone of intelligentization.
Is Huawei "Not Easily Replaceable"? Facing the homogenization anxiety brought by Huawei's "publicization," many automakers have begun to re-evaluate the selection space for intelligent driving solutions. Suppliers such as Momenta and Horizon Robotics have long coexisted with Huawei in the market, each serving the demand scenarios of different automakers and jointly forming a diversified ecosystem in the intelligent driving industry.
The rise of Momenta can be called a microcosm of China's intelligent driving suppliers. Founded in 2016, relying on its "Flywheel Large Model" technical architecture, the company has established partnerships with 9 of the world's top 10 OEMs, including Mercedes-Benz, BMW, Toyota, Honda, and General Motors.
At the 2026 Beijing Auto Show, Momenta CEO Cao Xudong revealed that its cumulative designated car models have exceeded 200, and the scale of equipped vehicles has surpassed 800,000 units. The launch of Mercedes-Benz's new all-electric GLC, equipped with Momenta's reinforcement learning large model, is even regarded as a landmark event of "German luxury standards + Chinese AI (Artificial Intelligence) technology."
Horizon Robotics is also gaining momentum. Its Journey 6 series chips have been designated by over 10 automakers, including SAIC Motor, Volkswagen, BYD, Li Auto, and GAC. In 2025, its mass production shipments crossed 10 million units, making it the first domestic intelligent driving technology brand to break through the 10-million-unit shipment mark. The intelligent driving solution developed by Carizon, the joint venture between Volkswagen and Horizon Robotics, based on the Journey 6M has been mass-produced on the Yu 06 and Yu 07 in May 2026.
The choice of different paths reflects the differentiated landscape of the intelligent driving industry. Momenta excels in algorithms and data flywheels, accumulating mass production experience by serving global leading automakers; Horizon Robotics focuses deeply on chips and toolchains, adapting to intelligent driving needs at different levels with an open architecture.
Huawei, on the other hand, relies on the native integration of the HarmonyOS ecosystem, mobile terminals, and cloud computing power to form a unique layout of cross-terminal collaboration. This all-scenario interconnection capability gives it a hard-to-replicate differentiated advantage at the ecosystem integration level.
The automakers pursuing in-house development present another picture. XPeng invests RMB 3.5 billion annually in intelligent driving, and in 2026, it plans to increase its R&D investment related to physical AI to RMB 7 billion. Its in-house Turing AI chip has achieved mass production and vehicle installation, showing unique advantages in the collaborative optimization of algorithms and hardware. NIO and Li Auto also have their respective layouts, mastering the discourse power of core technologies through vertical integration.
However, the in-house development route also faces realistic challenges: the cumulative shipment volume of chips is limited, making it difficult to form economies of scale in the short term; high R&D costs and fierce market competition create cost pressure, and the continuous operation of the data closed-loop also requires long-term investment. The strategic value of "full-stack in-house development" has been verified, but its commercial sustainability remains to be further tested by time.
What Zeng Qinglin meant by Huawei being "not easily replaceable" lies precisely in its composite barriers in technology, ecosystem, and brand. Jin Yuzhi pointed out at the 2026 Huawei Qiankun Media Day that the cumulative installed volume of Huawei Qiankun ADS has exceeded 1.9 million units, and the cumulative assisted driving mileage has even surpassed 13.7 billion kilometers. In this highly penetrated market, this is not a victory of a single technical parameter, but a moat built jointly by user scale, massive data accumulation, and brand trust.
Ultimately, this discussion about "replacement" cannot bypass a more essential proposition: when software-defined vehicles become a consensus, what kind of software-hardware collaboration does China's automotive industry really need? What are we afraid of losing, technology or the "soul"?
Finding the Balance Point Answering this question can be traced back to the "Soul Theory" that was fiercely discussed in the public opinion arena five years ago, and recalibrating its coordinates in today's industrial reality.
On June 30, 2021, then-Chairman Chen Hong, facing investors' questions about whether to cooperate with Huawei, threw out those words that were later repeatedly cited by the industry: "It is hard for SAIC Motor to accept a single supplier providing us with an overall solution. In this way, it would become the soul, and SAIC Motor would become the body. SAIC Motor cannot accept such a result; our soul must be held in our own hands."
This statement was later called the "Soul Theory," sparking widespread discussion in the industry. Five years later today, the cooperation models between automakers and Huawei have diverged: some companies choose deep binding, while others seek more initiative in the cooperation. The core proposition pointed to by the "Soul Theory"—how automakers can maintain autonomy and discourse power when cooperating with tech giants in the era of intelligentization—remains an unavoidable question for the industry.
Even more dramatic is SAIC Motor's own turn. In February 2025, SAIC Motor officially announced deep cooperation with Huawei, launching the fifth "Realm" under Harmony Intelligent Mobility Alliance (HIMA)—Shangjie. It is reported that Jia Jianxu, President of SAIC Motor, once consulted the retired Chen Hong about the intention to cooperate, and the latter responded: "Huawei has such great momentum now; you should go and cooperate with them." From "rejection" to "embrace," SAIC Motor's turn is not a denial of its past stance, but a pragmatic return to industrial laws.
The so-called "Soul Theory" can be deconstructed into the following three layers, which may help us clarify the logic behind SAIC Motor's turn. The first layer is hardware—the physical integration of chips, sensors, and actuators. This layer is already highly standardized, and suppliers like Horizon Robotics and Hesai are fully competent. The second layer is software—operating systems, algorithms, middleware, etc., where Momenta, DJI, and others are rapidly catching up. Only the third layer, the ecosystem—cross-terminal flow of user data, account systems, and service scenarios, as well as the user mindset that "a certain brand equals intelligence"—remains irreplaceable for Huawei.
"Bosch and Continental have been suppliers for many years; why is no one discussing the 'soul'? Because they are rooted in the mechanical and electronic control layers, which are relatively far from user perception," the aforementioned analyst pointed out. "Huawei, however, enters the core domain of intelligentization, directly defining human-machine interaction and user experience, making it naturally more visible and discussable."
Analysts from Gasgoo believe that although Huawei has clearly positioned itself as "not manufacturing cars," its actual actions have exceeded the scope of traditional Tier 1 suppliers. "No traditional component manufacturer would hold press conferences for automakers or drive traffic through channels, but Huawei did." It was also stated, "From the consumers' perspective, what they are buying is 'Huawei's car'." This shift in brand cognition is perhaps the biggest difference between Huawei and traditional suppliers.
It was further pointed out: "Before Huawei entered the automotive sector, consumers had already come to recognize this 'soul'. This 'soul' now extends to one additional product line: automobiles."
Ultimately, chips can be swapped, algorithms can be bought from multiple sources, but what truly makes Huawei inescapable is its already established ecosystem—user data, account systems, cross-terminal flow, coupled with the market cognition that "Huawei equals intelligence." These are not single technologies, but barriers stacked layer by layer.
The question today lies only in: at what cost are automakers willing to exchange for this layer, and whether they can hold their own boundaries.
Facing the realistic industry anxiety, Huawei itself is also adjusting. Jin Yuzhi reiterated the "electronic screw" positioning at the 2026 Huawei Qiankun Media Day, clarifying the parallel operation of dual models: full-stack cooperation and component supply. This "decentralized" strategy has, to a certain extent, responded to automakers' demands for differentiation. When the deep binding model encounters doubts about the "Huawei tax" and anxiety over brand independence, the "screw" positioning may provide a "way out."
Conclusion: As Huawei's automotive intelligentization business penetrates deeper, some automakers inevitably have to answer an old question: Without Huawei, who am I?
Some pursue full-stack in-house development, some embrace openness, and some fight with both hands... The paths vary, but they all point to the same essence: technology can be bought, ecosystems are hard to replicate, and brand recognition takes time to build.
There is no reference answer for this path, but those who only know how to copy others' homework are highly unlikely to go far.