On September 16, 2026, Infineon and United Automotive Electronic Systems (UAES) officially unveiled the plaque for the "Innovation Application Center (Chongqing)". The official press release was written with restraint, giving the impression: we have just opened another office in China.
However, those who understand the underlying strategy will feel a chill: for domestic peers, this "office" is like yet another nail driven straight into the "comfort zone" of domestic substitution.
The nail hammered in Shanghai ten years ago has grown into the first generation of zone controllers, multi-in-one motor controllers, local packaging and testing, and a host of vehicle platforms; the one being driven in Chongqing aims to pierce through the next-generation vehicle architecture, the automotive-grade heart of AURIX TC4x, the full-stack 48V power supply, the platformization of on-board charging, and that seemingly most inconspicuous yet most critical phrase: "digital platform and verification cycle compression".
Breaking down these moves, you will find that what Infineon has always wanted is not simply to sell a few more chips.
Because selling chips is a business, but defining the development chain is power.
Over the past decade, the most valuable output of the Shanghai center is a repeatedly validated working method: seating Infineon's chip engineers right next to UAES's system engineers, starting to discuss one thing before the OEMs have even finalized the BOM, locked the architecture, or finished writing the requirement documents:
What regulations your vehicle needs to pass next year, how much system cost needs to be reduced the year after, and how much margin for functional safety and cybersecurity needs to be reserved when going global to Europe, the Middle East, and South America the year after that.
This step is called "front-loading". Whoever enters the requirement phase first becomes the teacher; whoever only quotes prices at the front end of mass production, sorry, is just a backup. It is this exact move that allowed Infineon to spend ten years evolving from a "component supplier" into an "early development partner for China's smart vehicles".
*Introduction to UAES
UAES is not a randomly chosen partner. The 30-year joint venture foundation between Bosch and SAIC has equipped it with an understanding of the German functional safety paradigm, while immersing it in China's fiercely competitive NEV market to hone its speed. Infineon chose it for a shared language: chip makers fear two things most—understanding silicon but not vehicles, and understanding vehicles but not manufacturing.
UAES can translate AURIX, power devices, and power links into how motors torque, how braking is distributed, how domain controllers are decoupled, and how OTA is rolled back. Thus, what Shanghai secured ten years ago was the first ticket to China's domain control era; what Chongqing aims to secure over the next ten years is a more expensive and deeper ticket to software-defined vehicles, 48V mild-hybrid and hybrid reconstruction, central computing plus zone control, and on-board charging platformization.
Looking at the Chongqing location again, the layout is exquisitely strategic.
Changan and Seres are rooted by the river. Chengdu-Chongqing, linked with Wuhan, Hefei, and Xi'an, forms a "new automotive waistline". The discourse power of extended-range and hybrid technologies, road testing scenarios for intelligent driving, and the density of OEMs are all twisted into a single rope along this line.
In the past, the China headquarters of foreign-invested semiconductors were locked in Shanghai to serve joint-venture OEMs and the Yangtze River Delta supply chain; today, half of the "engines" of China's smart EVs are in the Yangtze River Delta, and the other half are right on this waistline.
By bringing the Innovation Application Center into the UAES Chongqing branch, Infineon has effectively grid-connected three capabilities: German chip architecture, UAES system integration, and western OEM scenario applications.
Therefore, an interesting scenario may emerge in the future: arguing over architecture with OEMs in Chongqing from Monday to Friday, and returning to Shanghai on weekends to align roadmaps with global R&D. While domestic chip companies might not necessarily fear this pace, international peers still relying on distributors to cover the field will be directly left out of the development window.
Now let's look at the technology checklist.
AURIX TC4x is the trump card in Infineon's hand for the new generation of automotive-grade MCUs. Featuring 28nm process, continuation of the TriCore architecture, parallel processing units, Ethernet, CAN XL, and PCIe all in one, it is essentially the control heart prepared for SDVs and domain control.
Infineon has not separately disclosed the market share of TC3x in China; however, it officially announced that its global automotive MCU share has reached 36%, and TC3x is precisely the mainstream choice for Chinese OEMs and Tier 1 suppliers like UAES when building zone controllers and powertrain domains. What TC4x aims to achieve is not simply "replacing TC3x"; its intention is to transform Infineon from a "reliable old partner" into the "default foundation for the next-generation E/E architecture".
What is a default foundation? It means that when OEMs write the architecture, they no longer discuss whether to use it, but only how to use it. The Chongqing center talks about "reconstruction of the embedded control core", which translates to: do not wait for OEMs to finalize the architecture before selecting the MCU; let's define the memory mapping, functional safety island, network topology, and underlying software middleware of TC4x all together right now.
In other words, whoever defines the architectural semantics defines the supply chain for the next decade.
The full-stack power supply for the 48V vehicle architecture is even more worthy of study. Many in the industry still think 48V is just for mild hybrids to save some fuel. In reality, in new-generation hybrids, extended-range EVs, steer-by-wire chassis, high-power accessories, and central domain power supply, 48V is the entry point for upgrading from an "electrical appendage" to an "energy hub".
Infineon holds a complete arsenal of power semiconductors, PMICs, MOSFETs, SiC, drivers, and sampling components. What it wants to sell is a power supply worldview "from the battery pack to the domain control board, and from the OBC to the electric drive auxiliary power source".
The platformization of on-board charging is not just about helping UAES draw a schematic; it is about turning the OBC, DC-DC, PDU, 48V power supply, and low-voltage energy storage into reusable Lego blocks. Once platformization is defined first by foreign enterprises and top Tier 1 suppliers, latecomer chip companies can only do pin-to-pin replacements; profits and discourse power will not be on the same level at all.
However, the most hidden and critical aspect is the "verification cycle". The craziest thing about Chinese OEMs is not their high configurations, but their fast pace: a major facelift every year, a new platform every 18 months, and software iterations every two weeks. The nature of automotive-grade chips is conservative: slow certification, slow PPAP, slow AEC-Q, and slow functional safety validation. By the time you finish the process, the vehicle model has already been facelifted three times.
When Infineon and UAES discuss expanding digital platforms and compressing the verification cycle for motor and domain control, they are performing a bridging act: using virtual verification, early models, joint simulation, and shared requirement baselines to compress the original 18-month component introduction cycle to sync with the vehicle model pace. Once this step is successfully executed, Infineon will no longer be a "high quality but slow" German factory, but will transform into a hybrid species with "German quality and Shenzhen speed".
Therefore, the most formidable foreign enterprise is never the one just shouting localization slogans, but the one that blends its home country's engineering discipline with the wolf-like aggressiveness of China's project-based system.
*Timeline of Cooperation Between Infineon and UAES
Some might ask, given how fiercely competitive domestic substitution has become, is this big brother Infineon still holding up?
I think we should really look past sentiment and focus on positioning.
It is true that domestic MCUs, power devices, and sensor chips are naturally surging forward, the RISC-V automotive-grade ecosystem is taking shape, and the full-chain localization of SiC is also accelerating.
But automotive semiconductors are not like shelves on Taobao where the cheapest wins; it is a game of "architectural power".
Infineon's strategy is exceptionally shrewd:
First, it does not engage in price wars with domestic chips, but moves towards the early stages of the system, ensuring that OEMs are already accustomed to its safety models, memory structures, power topologies, toolchains, and AUTOSAR ecosystem when writing requirements;
Second, it does not treat China merely as a sales market, but as a product definition market. Local front-end and back-end processes for TC3x, local production of TC4x, advanced mass production of 40V MOSFETs in Wuxi, and localization of radar sensors are all about "In China, for China", and even "From China, to the world";
Third, it never binds itself to a single OEM, but co-builds Innovation Application Centers with over a dozen top OEMs and Tier 1 suppliers, weaving a "semi-official yet non-equity" collaboration network that is more flexible than equity investment and deeper than distribution.
This approach is called "Stakeholder Tiefenbindung" (deep stakeholder binding) in German; in our village, we call it: I may not necessarily hold a controlling stake in you, but I will embed myself so deeply that you can't get rid of me.
Infineon dissects its own capabilities into a whole chain of components, architecture, tools, safety, power, training, logistics, and packaging/testing, and then stitches them respectively into the R&D, procurement, quality, software, platform, and overseas compliance of Chinese OEMs.
It is easy for you to change the MCU, but hard to replace the entire methodology "from chips to domain control to power supply to verification to overseas certification". This is the gentlest yet deadliest strike against domestic semiconductors: no need to kick you off the table, but keep you sitting on the edge of the table long-term as a backup, for cost reduction, and as a second source, while the core architectural position is always reserved for me.
Zooming out further. After 2026, the sovereignty of the automotive industry will no longer belong solely to OEMs, nor solely to battery manufacturers, but to those who can define the E/E architecture.
Tesla has re-taught the world with central computing plus zone control; Chinese OEMs have pushed the architecture to a new generation every three years using extended-range, urban NOA, 800V, 48V, and cabin-driving integration; European OEMs are still making slow turns within functional safety and compliance; and Japan is hesitating while guarding its hybrid moat.
Infineon, as a banner of European semiconductors, has made a very "non-European" move in China: moving the decision chain, application chain, and verification chain to the customer site.
Because it knows better than anyone that if it continues to define products in Munich, package them in Wuxi, and distribute them through agency channels, it will degenerate into a "rent collector of the previous generation's electrification dividend" in the SDV era.
Therefore, it learns from Chinese companies to form iron triangles, create joint war zones, and engage in scenario co-creation, while tightly holding onto the two most valuable assets of a German factory: the belief in automotive-grade quality and the underlying capability in functional safety.
This is why the Chongqing center is worthy of study. It is not just an office, but the "western interface" of Infineon's China strategy: connecting eastward to the decade-long assets of the Shanghai headquarters, westward to the war rooms of Chengdu-Chongqing OEMs, downward to the mass production lines of UAES's motor and domain control, and upward to future concepts like AURIX, 48V, OBC, zone power supply, and software-defined power.
In the next three years, if a new architecture vehicle model from Changan, Seres, Chery, Geely, or even the BYD system releases a jointly defined zone controller reference design by "Infineon plus UAES", there is no need to be surprised—that is not a supplier winning a bid, but an architectural marriage.
Therefore, there is reason to predict that Infineon's "Innovation Application Center Network" is highly likely to grow into an invisible operating system at the underlying layer of China's automotive semiconductor ecosystem.
It does not own OEMs, does not hold controlling stakes in Tier 1 suppliers, and does not compete for whole vehicle brands, but it holds a pen on the whiteboard of every key vehicle model project. OEMs think they are selecting chips, but in reality, they are selecting who can help them build cars on time, in compliance, mass-producible, and ready for global export. If technical communication upgrades from "quotes and parameters" to "how to build the next-generation vehicle architecture", the supplier identity is elevated to a co-author.
You see, what Infineon has always wanted is never just a single order, but to become a contributor to the underlying grammar of China's smart mobility industry.
*Infineon's Innovation Application Center Layout in China
So the key to this matter is front-loading. The decade in Shanghai proved that a chip maker can walk into the study of Tier 1 suppliers; the new game in Chongqing proves it can also walk into the battlefield of western OEMs. From selling chips to providing solutions; from local support to joint definition; from selling in China to defining in China and feeding back to the world... Infineon has completely rewritten the formula for the survival of foreign-invested semiconductors in China: I am not just a guest, but also a partner; not just a backup, but also the foundation; not just talking about quality, but also pace; not just protecting German genes, but also growing Chinese roots.
Of course, this game is not without variables.
Domestic chips will not stay in the substitution position forever. RISC-V, edge-side AI, cabin-driving integration, in-vehicle optical communication networks, and new power semiconductors are all redrawing the boundaries; nor will OEMs forever hand over architectural power. BYD, Huawei ecosystem, Xiaomi, and Li Auto are all pulling chips and systems into their own arms.
*The Chinese OEMs closest to full-stack internalization of "Chip—System—Software—Vehicle Model": BYD (Manufacturing + Power + Whole Vehicle Intelligence), Huawei (System + Chip + Ecosystem), NIO, XPeng, and Li Auto (Rapid convergence of ADAS chips + OS + Models).
No matter how shrewd Infineon is, it must answer a question: when more and more Chinese OEMs internalize the full stack of "chip—system—software—vehicle model" like Tesla, will foreign-invested semiconductors continue to be "high-grade cement", or retreat to "key material suppliers"?
The Chongqing center might just be its answer: relax, I am not here to steal your car, I am here to help you build it so well that you wouldn't dare to easily replace me.
Ultimately, what Infineon has hung in Chongqing is not just a plaque, but an anchor.
One end is tied to UAES's 30 years of system expertise, the other end is tied to the impulse of western OEMs for next-generation architecture, and in the middle are strung AURIX, 48V, OBC, domain control, verification platforms, and overseas compliance.
The competition in China's automotive semiconductors will ultimately no longer be such a naive dichotomy as "domestic versus imported", but will evolve into "who enters the development chain earlier, who penetrates the system layer deeper, and who can better weld the rigor of automotive standards with Chinese speed into one".
This move by Infineon seems gentle but is actually fierce; it seems like it is just pacing along, but in reality, it is grabbing the starting gun.
Ten years in Shanghai, another ten years in Chongqing. In the study of UAES, Infineon is quietly planning to write the underlying grammar of China's smart vehicles.
References:
[1] Infineon
[2] UAES