It’s being called “China speed,” defined by the accelerated rate at which software-defined vehicles can be designed, manufactured, and updated with new features. And nowhere is this hitting harder and forcing more profound changes than in Germany, Europe’s leading automotive market.

Rather than relying solely on customized electronic control units, SDVs use a combination of specialized and generic processors and memories wherever it makes sense to embed functionality into software. Tesla was one of the pioneers of this approach, but it has snowballed ever since Chinese automakers first entered the market in the early 2000s. Since then, even latecomers have managed to pick up share and speed. Xiaoping Motors (XPENG), founded in 2014, introduced its first vehicle in just four years. It delivered 190,000 vehicles in 2024 [1]. Li Auto, founded in 2015, introduced its first vehicle in 2019. It produced 501,000 vehicles in 2024.[2]

Collectively, China produced 31.3 million vehicles in 2024.[3] Germany, in contrast, manufactured just 4.1 million passenger cars in 2024.[4]

“The Asian OEMs are much faster in getting a car on the street,” said Roland Jancke, head of department for Design Methodology at Fraunhofer IIS/EAS. “The question is how we, as Western Europe or the Western world, can keep up with that speed. How can we get faster? One of the answers is automotive systems engineering. This is exactly the answer to speed up the development process. The automotive development cycle is still seven years from the start of the project until the start of production, due to the old development processes. But there has to be more speed. It has to get faster. What does automotive systems engineering mean? It means that the overall development process is driven by a model of the overall project. You start at the very beginning with model-based development, which is a different approach than is done today. Today, when a project is started, it is first split up into several sub-projects and this is given to different teams who don’t know what the others are doing, or don’t have too much interaction with each other.”

Germany is still the European powerhouse for automotive, but concerns about its global competitiveness stem from its traditional role as a producer of hardware-defined vehicles. OEMs like Mercedes-Benz, Volkswagen, and BMW have a standard way of operating and building out hardware-defined vehicles, typically taking years to test and produce new models, said Hanno Wolff, executive director of automotive business development at Synopsys.

Those OEMs that have shifted into SDVs have done so via workforce development, building out software departments within existing organizational structures, or through collaboration with Chinese companies. For example, Volkswagen partnered with XPENG to introduce China Electrical Architecture, a zonal electrical/electronic architecture.

What exactly is software-defined hardware? Software-defined hardware doesn’t mean different functions are actually written in software. Even though running software on a generic processor adds greater flexibility and shorter time-to-market, it’s slower than customized silicon, not as energy-efficient, and less secure.

“When the feature is hampered by latency or the computational aspect of the software side, that’s where you draw the line,” said Mike Yeager, senior vice president and general manager for Ethernet solutions at Infineon. “It’s not always clear where that delineation is. But if the computational cycles that occur from compiling and running software are prohibitive, or noticeable by the user, it’s not going to work. The software definition is there. It’s just what gets put into hardware may be different. The combination of the two really is what works best. You can still program it in, and you can do over-the-air updates, but you’d better be empowering the hardware to do a lot of the work.”

Fig. 1: SDV-based zonal architectures with MCUs and SoCs. Source: Infineon

Others agree. “The way we look at it is that it’s more about enabling an infrastructure when it comes to automotive, pushing software updates and new feature upgrades even after the vehicle goes out of the dealership,” said Siraj Gajendra, vice president of automotive products for software and solutions at Arm. “But when it comes to software-defined hardware core design, the fundamental aim is to ensure that the hardware being developed is aware of what software application it needs to run. At the same time, there is a lot more that is purpose-built.”

Co-design of hardware and software has been around for some time, and so have discussions about which comes first. But established automotive companies have largely stuck to their traditional approaches because there was no immediate competitive reason to make a costly change.

“Software-defined includes several things that are different,” said David Fritz, vice president of hybrid physical and virtual systems at Siemens EDA. “The software has to be able to impact hardware, and hardware needs to have an impact on the software. It’s that co-dependency that you need to exploit if you really want to have a differentiated product. There is a whole infrastructure that needs to be built around the hardware and the software team to enable that. One of the key points there is, ‘How do you actually validate that you’re headed toward a solution that’s going to meet your requirements down the road? No longer can we do this via a spreadsheet, or just do iteration generation after generation. It all has to be designed to purpose.”

Fig. 2: Key architectural directions for SDV. Source: Siemens EDA

**SDV partnerships ** This is true for all traditional carmakers, but it’s been easier for startup EV companies because they don’t have the legacy infrastructure. There are roughly 200 parts in an EV motor, versus more than 2,000 in an internal combustion engine (ICE), and customer focus is on electronics they directly interface with, such as the infotainment system, rather than the mechanical features in an ICE vehicle. In an EV, acceleration is already substantial enough for many consumers because of the high torque.

The challenge for traditional automakers is that they still need to maintain their ICE operations, while also investing in an EV infrastructure that includes software development teams and software-driven options. They also must compete against direct investments by the Chinese government, depreciated capacity by companies like Tesla, as well as lower labor costs in automated factories.

“I went to a conference recently, and there was a presentation about what different regions value the most in innovation in vehicles,” said Rob Fisher, senior director of product management at Imagination Technologies. “There’s a stark difference between the European market, which values safety and quality, and some of the more safety-related, ADAS-type features, whereas in China specifically the market is driven by in-car entertainment, voice interfaces, and interactions in the cabin. So there’s very different market needs there. I’m not personally convinced that the German market will follow the China market in the way they execute on things, which is a much quicker pace of development, just because the German market focuses more on the European market, which has different needs.”

Besides demographic-based differences in automotive feature preferences, there are also a host of challenges facing German carmakers in producing SDVs.

“A whole series of wakeup calls have been received by the German car industry, one worse than the next,” said Christian Burisch, strategic partnerships manager at Imagination Technologies. “They know that things have to change, but the problems that they’re all struggling with are more organizational. It is kind of hard if you’ve got 50 years of legacy of doing something a certain way, and suddenly you have to do it a different way. That change can be very hard.”

On the positive side, Germany has the potential to leverage local resources, including research institutions and chipmakers, to build out local semiconductor capacity for SDVs.

“There are multiple programs involving German players or institutions. Infineon, with its AURIX TC4xx family, supplies MCUs for zonal controllers,” said Pierrick Boulay, principal analyst for automotive semiconductors at Yole Group. “ESMC foundry (TSMC 70%, Bosch/Infineon/NXP 10% each) is aimed at automotive semiconductors, but it will cover any type of semiconductors, not only those related to SDVs. GlobalFoundries, with its capacity expansion in Dresden, should cover any type of semiconductors used in cars.”

Leveraging this advantage would require coordination and a reworking of long-held processes, such as the way OEMs historically have purchased components necessary for vehicle designs.

“All the car companies are very much geared toward buying the cheapest component that serves their needs in a particular model, and they have whole departments geared toward doing this,” said Imagination’s Burisch. “But if you have a software-defined vehicle, that means that you have to be able to add software value to it afterwards, which means you have to over-spec your hardware, which is something that does not come naturally. And they’re not organizationally set up to do it, because there are departments that are being rewarded for making sure they get the cheapest thing that serves that particular need.”

A lot of the norms involved in traditional OEMs contrast starkly with how Chinese startups are being run, with software experts at the helm. “If you look at the CVs of C-level executives at German OEMs, you’re going to find a similar pedigree,” said Synopsys’ Wolff. “They studied in Stuttgart, Munich, or Brunswick, and they got to where they are because they’re mechanical engineers. They understand combustion engines.”

Eliminating what a company does best, especially from the top-down, means replacing it with something better for the consumer. In this case, replacing or adding hardware expertise with software would mean creating value for an OEM by developing an SoC or software stack.

“OEMs must first decide whether financially it’s viable,” said Fisher. “Step two is figuring out how they execute on that and how they build that capability.”

Organizational and cultural challenges German OEMs, as well as their counterparts in Europe and the U.S., historically were accustomed to a 7-year design cycle. With a software-defined SoC or software stack, they need to develop it much more quickly and embed it into vehicle architecture to remain competitive. As vehicles become updatable and data-driven, Germany’s automotive industry must rework its approach to hardware–software collaboration, both culturally and organizationally.

“To be future-ready, automotive companies must integrate software leadership into their traditionally hardware-centric organizations,” said Frank Bösenberg, managing director at Silicon Saxony, a microelectronics and information and communication cluster in Germany. “Bosch and Mercedes have made an early start with integration. The next step is to completely rewire the organization, moving from hierarchical engineering to agile, cross-domain product teams that jointly own both the code and the hardware lifecycle. The key to success will be merging safety-critical rigor with software agility — integrating DevSecOps principles into the industrial engineering process while maintaining Germany’s quality standards.”

Bosch, for example, wants to turn cars into personal assistants, noting that the future of vehicles means hardware being designed to fit software. The company laid out a plan at IAA Mobility 2025 for a software-driven future, partnering with autonomous driving leaders, and highlighting its ADAS product family.

Mercedes-Benz, meanwhile, introduced its MB.OS operating system, which is a chip-to-cloud architecture. MB.OS is launching with Mercedes’ newest CLA model, and will gradually roll out across the entire portfolio. “MB.OS is a data-supported and flexibly updatable operating system that connects all major vehicle domains — infotainment, automated driving, body and comfort, and driving and charging. It forms the digital interface to our customers while ensuring robust data protection and compliance with global standards,” a spokesperson from Mercedes-Benz said.

The company created 3,000 new jobs in software development to implement MB.OS, 1,000 of which are based out of the central development and training hub in Sindelfingen. “Software development for MB.OS is a global effort, with teams working across hubs in Sindelfingen, Berlin, Tel Aviv, California, Beijing, Shanghai, Kawasaki, Seoul, and Bangalore. Through our ‘Operational Excellence’ initiative, we’re scaling agile models to maximize customer value and accelerate innovation,” the spokesperson said.

The effort is part of a “digital first” approach similar to what Bosch has outlined.

“Germany’s strength in mechanical precision must now evolve into mastery at a systems level,” Silicon Saxony’s Bösenberg said. “The challenge lies not in the linkage between software and hardware, but in cultural and operational speed. Traditional governance structures and siloed development still hinder the convergence of hardware, software and data-driven architectures. To succeed in this transformation, Germany must rethink its engineering DNA and integrate semiconductor, software, and AI development cycles under unified, agile frameworks.”

The trajectories of German car manufacturers are diverging. Mercedes’ in-house operating system strategy signals progress, but true competitiveness will depend on whether OEMs master agile release cycles, ensuring software safety and embedding continuous innovation into product lifecycles.

“The next frontier is platformization — the ability to integrate vehicle software, edge computing and cloud analytics to create scalable, data-driven business models,” he said. “Tesla and leading Chinese OEMs currently set the pace in this domain, and Europe must close the execution gap.”

German OEMs are working to close that gap, albeit at different rates. According to an executive guide released by Synopsys, the automotive industry is undergoing its most profound transformation in over a century.

Others agree. “We’ve been in an era of you design the hardware first, and then you write software for it, and that’s done us very well for a long time,” said Alex Starr, corporate fellow at AMD. “But we’re seeing new techniques coming out. Architecture is changing from a software perspective, and that is creating more flexibility. That hardware needs to adapt and become more flexible. It’s ultimately some kind of tradeoff between how much bespoke hardware you have, and how flexible that is. How flexible is your software? We’re in this place now where we’re turning the dial a little more toward the software side to deliberately create more flexibility. That creates challenges economically, but that’s generally how we see things getting more to a software-first approach.”

Production, pace and partnerships To some, a collaborative approach with SDV startups is the path forward for old-school OEMs that can’t quite get up to speed alone.

“German chipmakers, such as Infineon, Bosch, X-FAB, Elmos, plus some fabs located in Germany, are traditionally strong in automotive semiconductors, which are still reinforcing their position,” said Yu Yang, Principal Analyst, Automotive Semiconductors at Yole Group. “The challenge lies in the lack of advanced nodes, both in design and manufacturing. These segments, mostly stemming from consumer electronics business — for example Intel, NVIDIA, Qualcomm, AMD, etc. — require huge investments. There has been a rise of in-house designs of advanced SoC chips in the automotive supply chain, as led by Tesla and quickly followed by Chinese car makers, such as XPENG, Nio, and others that have not taped out yet. However, we don’t see any efforts locally in Europe, including Germany premium brands. From the manufacturing side, as Intel holds back its investment plan in Magdeburg for the 1nm node, the lack of advanced capacity in Europe is not fixed. There is only a 3nm Intel fab in Ireland, but not covering automotive applications. And other capacities, either existing or under construction, are far from the competing nodes of at least 7nm for advanced SoCs. This is not only a negative factor in SDV development, but also a risk in the supply chain.”

Bosch already is working with WeRide and Horizon Robotics on the development of systems for assisted and automated driving in China and in Europe with Volkswagen subsidiary Cariad. Bosch said alliances like these will become increasingly important in the future, and that the proportion of software in vehicles will continue to rise, along with the benefits for drivers in safety, convenience, and mobility.

OEM reorganizations are happening — VW, Bosch, and MB are proof of that, but the pace of change is still slow in Germany, as China continues to evolve and build out infrastructure that further widens the SDV gap.

“The Chinese are now creating or ramping up production facilities and their own supply chain with their own tier ones,” said Wolff. “We see, for instance, Renault developing in China for Europe. It’s selling almost no cars in China anymore, but they have R&D facilities in China. This is turning the whole thing around. And despite all the tariffs, BYD and others are coming heavily into the arena. So it’s a question of how fast the crash is happening in order to really substantially change things — and maybe another aspect to that [is] that there are 130 OEMs in China, and the plan is that there will be three to five global champions. If you look at the shareholding structures in German OEMs from China, how will that evolve eventually?”

Security and EU regulations and roadblocks European regulations are also evolving quickly, with new frameworks like the EU AI Act, the Cyber Resilience Act, ISO 21434, ISO 26262, and UNECE R155/R156, making certified security a strategic requirement instead of a technical option.

At the same time, the EU is losing ground in the global market, according to a study commissioned by the European Association of Automotive Suppliers. Suppliers face a cost disadvantage of 15% to 35%. Key drivers of the deficit include high energy and labor costs, regulatory burdens, and fragmented frameworks. In contrast, China and the U.S. combine industrial support measures with protective mechanisms, creating structural disadvantages and unfair competition.

“Edge AI and autonomous driving are accelerating the shift to software-defined vehicles, where most safety-critical decisions will be made in real-time at the vehicle edge,” said Adiel Bahrouch, director of business development at Rambus. “This makes the underlying compute platforms, AI models, and data pipelines high-value security targets.”

Rambus has a three-pillar strategy to address emerging threats (cybersecurity, supply chain, firmware attacks, etc.) related to SDV safety, including:

• Hardware root of trust;
• Certified, standards-based security, and
• Secure manufacturing and supply chain protection.

“One of the biggest emerging threats with Software Defined Vehicles is the rapid shift toward software-driven architecture,” said Bahrouch. “As SDVs rely predominantly on software, always connected to cloud, mobile devices, and V2X infrastructure, and continuously updated, the attack surface and direct entry points increase dramatically. Remote cyberattacks at scale, compromised OTA updates, and firmware manipulation are becoming high-risk scenarios. At the same time, decoupling software from hardware introduces more players and new market entrants in the supply chain, amplifying risks of firmware tampering, counterfeit components, and malicious code injection across the supply chain.”

Conclusion Germany’s success, as well as those of its EU neighbors, will hinge on its ability to convert engineering expertise into digital speed by building a sovereign European SDV stack that integrates semiconductors, cloud infrastructure and mobility software, while balancing compliance, resilience and scale.

“The defining inflection point will be Europe’s capacity to industrialize these technologies faster than geopolitical dependencies deepen. Failure would result in Europe becoming a premium integrator within someone else’s ecosystem, whereas success would secure its position as a global hub for safe, intelligent mobility systems,” Bösenberg said. “The central challenge is not recruitment, but identity. Automotive OEMs must redefine themselves as software-driven mobility companies. Their competitiveness will depend on hybrid talent models that combine expertise in embedded systems, AI engineering and semiconductors. Dual career paths, partnerships with research institutions and AI-assisted engineering environments will help to attract and retain the best digital talent, who want to work on scalable, mission-critical systems, not just cars.”

— Ed Sperling and Ann Mutschler contributed to this report.

References

1. NASDAQ
2. NASDAQ
3. CarNewChina.com
4. Germany Trade & Invest, Germany’s economic development agency.

Related Reading SDVs And AI Forcing Big Changes In Automotive Advancements in vehicle architectures, AI, and security are reshaping the future of the car. Automotive Outlook 2025: Ecosystem Pivots Around SDV OEMs and suppliers are beginning to move in lock-step, linking software design with chip development to speed time to market, reduce costs, and help future-proof vehicles.