Huawei is advising automakers to reconsider their approach to developing advanced driver assistance systems (ADAS), suggesting that not every component needs to be built in-house. According to a report by HUXIU, Jin Yuzhi, Huawei's senior vice president and CEO of Yinwang, a subsidiary focused on automotive technology, said in a September 28 interview that automakers should weigh the costs of self-development against purchasing from suppliers.
Jin highlighted that research and development (R&D) requires continuous investment, and if sales volumes cannot support the scale, self-development may not be cheaper than procurement. He advocated for a clearer division of labor between automakers and suppliers. The suggestion comes amid debate over whether Huawei is taking too much profit from automakers by selling its technology.
Huawei disclosed that its Qiankun ADAS cumulative installations surpassed 2 million units on August 27, 2026. The first million installations took 44 months, while the second million took just 12 months. The company also reported that its intelligent automotive solutions business generated 45.018 billion yuan (about $6.34 billion) in revenue in 2025, a 72.1% increase year-on-year, according to its annual report.
The cost of self-development
Jin provided a figure for Huawei's Qiankun R&D budget this year: it has been raised from 18 billion yuan at the start of the year to over 19 billion yuan (about $2.68 billion). He explained that mature systems require continuous iteration, with ongoing investment in software, manpower, computing power, hardware, and domain controllers, which do not end with the delivery of the first product.
This budget is not a direct quote for any automaker's self-development costs, as technical routes, feature scopes, and existing capabilities vary. However, it serves as a reminder that entering this business requires sustained resource allocation for subsequent iterations. The first investment solves whether a system can be built; subsequent investments determine whether it can keep up with evolving demands.
Scale becomes crucial. How many vehicles use a technology affects how R&D costs are amortized, and whether it can deliver product differentiation that consumers are willing to pay for influences returns. The economics of self-development depend on both ends.
For automakers, the bill also includes time. Once a development team is established, when the product can be delivered and whether it meets vehicle launch schedules will affect the outcome. Saving on procurement costs but delaying a key product may not be a wise trade-off.
Procurement, however, is not just about the contract price. Integrating a solution into a vehicle requires adaptation, validation, vehicle integration, and ongoing maintenance. Who bears these costs and how they are calculated must also be part of the comparison. The real comparison is the long-term cost of each route under similar functionality and delivery requirements.
Division of labor and the BYD example
Jin's proposed division of labor would have automakers focus on vehicle definition, chassis, vehicle integration, and user operations, while suppliers leverage their technological expertise. This requires automakers to reassess which capabilities are worth long-term investment and which can be obtained through cooperation.
Huawei's ability to take on these tasks depends on whether it can turn continuous investment into competitive products and pricing. The company's growing installation base suggests its technology is entering more production vehicles, allowing common R&D investments to be reused across projects. Each new cooperation model still requires adaptation and validation work, but scale can offer cost advantages depending on how much of the common technology can be reused.
A concrete example is BYD's cooperation with Huawei. In August 2024, BYD's Fangchengbao brand announced a smart driving cooperation agreement with Huawei, with the first product being the Bao 8. The technology combination includes Huawei's Qiankun ADAS 3.0, BYD's 2.0T DMO hybrid off-road platform, and the full-function YunNian-P intelligent hydraulic body control system. This shows that different technologies can serve a vehicle's positioning together, with the automaker integrating them into a user-perceivable experience.
Using the same supplier does not mean building the same car. Beyond assisted driving, space, energy consumption, chassis, design, price, and service still constitute purchase decisions. Automakers must differentiate through these elements, not just by announcing the adoption of a technology.
The recent adjustment of AITO, a brand under HarmonyOS, further illustrates this division. On September 15, HarmonyOS and AITO announced that AITO would remain within the HarmonyOS cooperation framework, with product definition, design, brand marketing, channel retail, and service system led by Seres, and Huawei Terminal participating in empowerment.
As more brands cooperate with Huawei, a question arises: how can automakers turn purchased technology into their own product advantage? While a technology can be a key selling point, it may not remain a brand-exclusive advantage if competitors also adopt it. This is the next challenge for automakers in the division of labor.
