Beijing Auto Show Observations: Intelligent Driving Is Shifting from a Selling Point to a Standard Feature, Supply Chain Moves to the Forefront
This year’s Beijing Auto Show is quite large in scale. Official data shows 1,451 vehicles on display, with 181 global debuts. The event is, of course, dominated by automakers, but the real excitement now goes beyond just the new cars themselves.
Two days before the official opening, Huawei held a smart automotive technology launch event in Beijing, showcasing 38 models using its solutions; on the same day, Horizon Robotics unveiled its cabin-driving fusion chip "Starry Sky" and its intelligent vehicle operating system; Bosch also brought their mass production timetable for advanced driver assistance, brake-by-wire, and steer-by-wire technology to the forefront. Before the exhibition halls officially opened, the supply chain had already taken center stage.
From Configuration to Production
This shift first manifests in a very specific product line. Over the past two years, urban navigation, LiDAR, and high-performance computing chips were mainly topics for high-end models — lots of spotlight at launch events, but real vehicle adoption was limited.
Now, at this year’s auto show, another set of questions is being asked repeatedly: How much can development cycles be shortened, how much can costs be reduced, and can a single system be reused across chips, models, and brands?
Reuters, citing Omdia’s assessment, said that in China’s market, new cars priced around 150,000 RMB are seeing advanced driver assistance and fast charging become increasingly close to standard features. With functions reaching this point, automakers are focusing less on novelty and more on delivery capability.
Horizon Robotics: A Mass Production Ledger
What Horizon Robotics presented this time is almost a mass production ledger.
According to the company’s disclosure on April 22, the "Starry Sky 6P" uses a 5-nanometer automotive-grade process, achieving 650 TOPS in computing power; more crucial are several following metrics: the cabin-driving fusion reduces vehicle space occupation by 50%, cuts integrated vehicle cost by 1,500 to 4,000 RMB per vehicle, and shortens R&D and delivery cycles from 18 months to 8 months. The partnership list already includes BYD, Volkswagen, Chery, Bosch, Denso, Valeo, and iCAR is named as the launch partner.
For automakers, this value can hardly be classified as "just another configuration item" anymore — it directly affects timetables, cost lines, vehicle architecture, and delivery pace.
WeRide: Reusability as the Key to Scale
The signals from WeRide take things a step further.
On April 24, the company announced WRD 3.0 now adapts to multiple mainstream in-car computing platforms, covering different gradients of computing power and cost; at the same time, it revealed that this solution has received nearly 30 model nominations from automakers such as GAC and Chery. A few days earlier, GAC Aion’s brand new N60 opened for pre-sale, with WeRide’s one-stage end-to-end assisted driving solution as standard on all variants.
As advanced smart driving moves into wider price brackets, platform migration and solution reusability become increasingly important. Often, what determines whether a system can be scaled up isn’t a new concept at a launch event, but whether automakers can avoid changing their architecture again or repeating validation steps.
Huawei: The Tech Platform Itself Is Already Generating Substantial Revenue
Huawei has taken a more straightforward approach: the technology platform itself is already capable of generating significant revenue.
Reuters reported on April 23 that Huawei will invest 18 billion RMB this year in smart driving R&D, with 10 billion RMB directed toward the computing power needed for training; in 2025, automotive-related revenue will grow by 72% year-on-year, reaching 45 billion RMB. Product coverage is rapidly expanding — among the 38 models displayed at this Beijing event, there were both Chinese and international brands, including Audi and Toyota.
When a supplier can put R&D investment, revenue growth, and model coverage on the table at the same time, the market naturally changes how it views them.
Bosch: The Execution Layer Steps Up
What really solidifies this line is the fact that the execution layer is starting to step up.
Bosch’s official press release in April was very detailed: from March 2026, the company will be allowed to conduct level 3 real-world road tests in Wuxi, with the system capable of autonomous braking, steering, acceleration, and lane changing up to 120 km/h; hydraulic brake-by-wire has signed supply contracts with five manufacturers, and is expected to enter passenger car mass production by mid-2026, whereas steer-by-wire will go into mass production on multiple platforms in China this year.
At this stage of advanced smart driving, it’s not just about "seeing" and "computing" — it’s about "making it happen." Without the closed loop at the execution layer — steering, braking, and chassis — chips, algorithms, and models at the front end can hardly be truly implemented.
Foreign Automakers: From Product Adaptation to Connecting with Local Technology Stack
The moves by foreign automakers make this point clear as well.
On April 21, Volkswagen released its products and artificial intelligence roadmap for China, with clear timelines: starting from 2026, new cars based on China’s electronic and electrical architecture will carry in-car intelligent models; by 2027, the next-generation architecture will integrate smart driving and cockpit control into a unified system. The ID. AURA T6 has been confirmed to use a locally developed China electronic and electrical architecture and CARIZON’s assisted driving solution. Meanwhile, among the models displayed by Huawei, there were also Audi and Toyota.
Foreign brands in China previously focused on product adaptation, but now look more like they are integrating with a full local technology stack.
The Layers Beneath the Main Stage Are Becoming Heavier
Putting together what these companies are doing, the main narrative at this year’s Beijing Auto Show becomes clear.
New cars will keep competing, prices will keep going lower, but the things that will set the pace for the next phase are increasingly being written into platforms. Whoever can connect chips, systems, training, deployment, and execution into a stable solution will more easily enter more models and brands, and also be better positioned to navigate price wars. The main stage is still centered on cars, but the layers underneath are gaining weight.
On the Ground, Focus on Three Segments
When this narrative turns to assets, the focus won’t be just any company associated with smart driving, but three more specific segments:
- Platform solution providers and chip companies: the market will watch if they can replicate across brands, models, and price segments;
- Leaders in the perception layer: especially those capable of moving from singular components to integrated solutions;
- Execution layer: brake-by-wire, steer-by-wire, and chassis domain control, which previously seemed like upgrades to components, will increasingly resemble advanced driver assistance infrastructure.
The differences here are not about who has the newer story, but about whose delivery increasingly looks like standard components that can be repeatedly called upon.
What to Watch in the Coming Quarters
In the coming quarters, the real focus will no longer be the spec sheets at launch events.
First, see whether the same platform can keep gaining nominations across brands; then, whether per-vehicle charging and software service business models become clearer; after that, watch the licensing speed for advanced driver assistance and whether execution layer components like brake-by-wire and steer-by-wire scale up as promised by manufacturers.
The Beijing Auto Show has pushed the supply chain to the foreground; what will determine how long it stays there is still delivery, orders, and mass production pace.
Risk Warning and DisclaimerThe market carries risks; investment should be prudent. This article does not constitute individual investment advice and does not take into account a particular user’s unique investment objectives, financial situation, or needs. Users should consider whether any opinions, views, or conclusions herein suit their specific circumstances. Investments based on this are at your own risk.