The automotive industry is currently experiencing a transformation of historic proportions: vehicles are increasingly becoming rolling software platforms - so-called Software-defined Vehicles (SDV). This development opens up enormous opportunities but also presents the industry with huge challenges. Ahead of the "Automotive Software Strategies" conference, we asked leading experts and members of the advisory board - Markus Blonn (IAV), Professor Dieter Nazareth (Landshut University), Andreas Herzig (Deloitte), Simon Fürst (BMW Group), and Joachim Langenwalter (TMT CoPilots) - about their views on technological trends and changes in the automotive world and what developments they predict.
China is increasingly perceived as a leading market for Software-defined Vehicles. What trends and challenges do you see here?
Markus Blonn, Senior Vice President at IAV, impressively describes how China is increasingly perceived as a driver of global automotive development. The speed with which new software functions are developed and implemented in China ("China-Speed") puts massive pressure on Western manufacturers. While the Chinese industry responds agilely to technological innovations, Western OEMs are slowed down by complex hardware architectures, processes focused on security rather than efficiency, and time-consuming committee structures. Blonn warns urgently: "Isn't the quality demand of Western manufacturers actually an expression of too much complexity in the hardware architecture?" He sees a critical shift of development capacities of Western manufacturers to China, which could limit the ability to adequately serve other markets.
Professor Dieter Nazareth from the University of Landshut adds to this assessment with cultural aspects. The car in China increasingly plays a completely different role than in the West: "While in the Western world driving is still the focus (joy of driving), in China it is the infotainment system of a vehicle (joy of being entertained)." In Chinese megacities, the vehicle becomes a rolling living room, as traffic conditions force longer stays.
Joachim Langenwalter, formerly responsible for automotive software and AI at Nvidia and as head of development at Stellantis and now working as an Executive Advisor, goes a step further: He sees China's strength not only in speed or agility, but in a fundamentally different approach. "In China, not only the vehicle is optimised, but the entire digital life of the user," he says. Using Xiaomi as an example, he describes how a vehicle seamlessly connects with the smart home, smartphone, and smart office - a level of comfort and connectivity that excites Chinese customers. "The Xiaomi vehicle sells better than some premium cars from German manufacturers because it is not just a car, but part of an ecosystem." While European OEMs continue to think functionally in silos, Chinese players consistently focus on user-centred overall solutions. Huawei is also working successfully with the Sares Group with the AITO brand, Changan, Avatr with its ecosystem. Additionally, the 996 work culture (9 am to 9 pm and 6 days a week) has become synonymous with the high pace in the Chinese tech industry, which we find hard to compensate for in Europe.
Andreas Herzig, Leader in Automotive Technology & Regulations at Deloitte, highlights that China and Europe have taken different technological paths. "While China has remained at Level 2 in recent years and developed Level 2+, we have moved strongly towards autonomy (Level 3, Level 3+)." Herzig sees the challenges for European manufacturers primarily in providing fast, flexible, and reliable software updates, which is not efficiently achieved in traditional organisations with established supply structures and the already described complex hardware architecture.
Simon Fürst, Cooperation Manager Automated Driving at the BMW Group, describes the technological division of the market even more clearly: "The RoW-OEMs (Editor's note: Rest of the world) have a long history in classic ADAS stacks and are now gradually switching to AI. BMW is the first car manufacturer worldwide to have approved the combination of the BMW Highway Assist (Level 2) and BMW Personal Pilot L3 (Level 3) driver assistance systems in a single vehicle - the BMW 7 Series. In contrast, China only entered ADAS with the AI wave, has no legacy, and has focused on an end-to-end customer experience from the outset. The HW and sensor equipment was often chosen generously according to the principle 'it will get cheaper anyway.'" Genuine Level 4 AD stacks only exist with robotaxi providers in the USA and China. Furthermore, in the area of ADAS, an increasing technological and geopolitical separation into "China" and "Rest of World" can be observed.
What developments do you expect in the field of Software-Defined Vehicles and what opportunities arise from them?
In the field of Software-Defined Vehicles, Markus Blonn expects a significant centralisation of electronic vehicle architectures, combined with a marked reduction in control units. Future software developments would be data-driven, highly collaborative, and AI-supported. This would enable a tremendous increase in development speed as well as entirely new business models, where software can be developed independently of specific hardware. Blonn sees clear advantages in this: "The reduction in hardware diversity leads to an increased value creation through software that is developed independently of hardware."
For Joachim Langenwalter, the shift to SDV means far more than a technological change - it demands a comprehensive paradigm shift: "Anyone who thinks they can simply implant SDV approaches into existing organisations is mistaken." From his perspective, the automotive industry needs changes on four levels: technology, corporate culture, business models, and customer relationships. The organisation must evolve from hardware variant sales to a systematic solution provider. He explains frequent failures with a vivid image: "Many OEMs build a speedboat - but crew it with the old team. That doesn't work." Only those who have the courage to truly think anew can fully exploit the potential of data-driven SDV architectures.
Professor Nazareth urgently warns that classic quality virtues must not be neglected in this process. He emphasises: "What use is the best infotainment system if the brakes are of poor quality?" This view is shared by other experts, including Andreas Herzig, who also sees the need to master regulatory requirements such as digital homologation, software asset management, and online accessibility of vehicles with sufficient bandwidth. Herzig specifies these challenges and explains that the reuse rate for software can be high, but only if it is possible to effectively address both organisational challenges such as asset management and regulatory requirements. He also emphasises that local developments ('local for local') are becoming increasingly important to reliably implement updates and functional extensions over the entire lifecycle of the vehicle in short cycles. Simon Fürst shares these views and additionally points out that "the biggest challenge is and remains that the traditional automotive industry (OEMs and Tier 1s) is still slow to transform into software-oriented companies."
What role does the integration of new technologies like RUST or AI play in the transformation of the automotive industry?
According to experts, the integration of RUST and AI will be crucial for the transformation in the automotive industry. Markus Blonn sees great potential for the programming language RUST, particularly in displacing C in embedded software development in the long term: "The advantages in terms of safe programming are immense for complex software and enable a permanent reduction in integration, testing, and debugging efforts." Additionally, developers simply enjoy and are interested in the Rust language.
Professor Nazareth, however, assesses RUST pragmatically and believes it plays a rather subordinate role, as it is ultimately just another tool to improve software quality. He explicitly emphasizes the enormous challenges that the integration of AI in safety-critical vehicle functions brings: "The introduction of AI in safety-critical functions is one of the major tasks. Mastering this is the decisive challenge of the future."
Andreas Herzig explains the possibilities of both technologies in a differentiated manner: "RUST is capable of avoiding common logic errors and goes significantly further than normal C or C++ compilers - a good approach to error prevention." In AI, he distinguishes between classical AI and generative AI. While generative AI could significantly improve the customer experience in the vehicle - for example, through voice and navigation-based assistants - the integration of classical AI in safety-relevant areas is much more complex: "In the area of prediction and planning, deterministic software is often still used, as proving functional safety and security is extremely difficult. However, I expect that this will increasingly come - risk and scenario-based testing in a virtual environment could be the solution here."
Simon Fürst clearly sees future development as AI-centred: "Future AD stacks will be entirely AI-based and will contain comparatively little traditional code. Safety will need to be demonstrated through extensive simulations and statistically." He adds that the automotive industry, due to its size and complexity, is unlikely to switch entirely to RUST. Nevertheless, he considers RUST to be useful when it is specifically used in areas with high ASIL (Automotive Safety Integrity Level).
Joachim Langenwalter also sees AI as the central lever of the SDV transformation - significantly ahead of new programming languages like RUST. These are certainly useful, for example in safety-critical applications, but the real lever lies in the availability and use of data. "The domains that will shape automotive manufacturing in the future - connectivity, autonomous driving, AI - come from computer science, not traditional engineering," he explains. This leads to massive cultural tensions: "In many companies, engineering careers still dominate, while tech talents are neither promoted nor retained." Those who do not open up here will lose touch - not with the technology, but with the people who develop it.
How can OEMs and suppliers collaborate more effectively to meet the challenges of modern software platforms?
Markus Blonn emphasises in response to this question that the traditional business and cooperation models have become obsolete due to the trend towards software-defined vehicles and calls for a fundamental redefinition: "The previous question of who does what, when, and where in the development process, and who earns money where, is no longer sufficient. It's about a paradigm shift in the entire value chain." Blonn recommends new, more integrated collaboration tools such as virtualisations, simulations, and cloud collaboration tools, as well as open-source projects for non-differentiating functions.
Professor Nazareth also calls for a shift from the traditional OEM-supplier relationship to a partnership model: "The classic OEM-supplier collaboration based on a specification sheet must give way to a collaborative business model." He sees cross-OEM standardisation, similar to AUTOSAR Classic, as essential for future software platforms.
Andreas Herzig adds to this perspective by highlighting the need to promote greater openness and transparency on both the OEM and supplier sides: "Constant manoeuvring by OEMs to squeeze a few more cents out of the price ultimately costs much more in quality costs. Suppliers must be willing, in return, to be much more open and to provide source code openly." Herzig sees a significant opportunity for companies in the use of open-source communities but also warns that companies that close themselves off could be pushed out of the market in the long term.
Langenwalter strikes the same chord and calls for a radical realignment of collaboration between OEMs and suppliers - moving away from short-term project awards towards data-based partnerships. "In the SDV era, I can no longer simply replace a supplier without jeopardising my update and software capabilities," he warns. Instead, stable relationships are needed throughout the entire vehicle and customer lifecycle - including shared data access, joint AI know-how, and a cooperatively developed data factory. The reality often looks different: "Many suppliers are cut off from crucial information - yet data is the foundation for everything that will happen in the vehicle in the future." For start-ups, this conflict means they must vertically integrate and handle differentiating architecture, hardware, software, and AI in-house, allowing them to act and react faster.
Simon Fürst clearly supports his colleagues' statements and predicts fundamental changes in collaboration and business models: "Collaboration will fundamentally change - white-box development and open source will dominate. Open-source software can be developed jointly by many users, including competitors." Fürst emphasises that such models are inevitable to meet the demands of future vehicle software. BMW is already taking this path in the Neue Klasse "where we consistently develop our automated driving and parking functions along the three pillars 'Smart, Safe, and Symbiotic' and thus create further added value tailored to specific market conditions for our customers."
Where is the journey heading for SDV?
What is the essence of this? Experts agree that the automotive industry is currently in the midst of a profound transformation. In the course of this development towards software-defined vehicles, established structures are being turned upside down, requiring significant adjustments from all stakeholders. China is increasingly becoming the lead market, particularly due to its rapid implementation speed and consistent customer orientation. At the same time, Western manufacturers are challenged to critically question and adapt their complex structures and quality standards to remain globally competitive.
New central vehicle architectures, accompanied by data-driven and AI-supported development processes, offer immense potential but also require profound changes in organization and collaboration. The programming language RUST and especially the use of artificial intelligence will significantly influence automotive development, even though their integration is associated with considerable challenges - particularly regarding safety and reliability.
The experts predict a fundamentally new way of collaboration between OEMs and suppliers, characterised by transparency, openness, and communal innovation. Business models will increasingly rely on collaborative, open-source-based approaches to meet the rising demands.
This article was first published at all-electronics.de
