AEK 2026: Why execution now matters more than strategy

The 30th AUTOMOBIL-ELEKTRONIK Kongress in Ludwigsburg should have been an anniversary celebration. Instead, it became a sober assessment of an industry under pressure. Between software-defined vehicles, AI-defined mobility, chiplets, edge AI, open source, semiconductor strategies, China-Speed and global supply chains, one message became clear: the automotive industry has identified its future topics, but the harder phase has only just begun.

Strategies now have to become production-ready products. Platform promises must turn into robust architectures. Calls for collaboration need to become working ecosystems. That made the 30th AUTOMOBIL-ELEKTRONIK Kongress less a nostalgic retrospective than a technical reality check.

Christian Sobottka, President and CEO of Harman, captured the mood with a question later picked up again by AEK Chairman Alfred Vollmer: does the industry want to be a victim, or does it want to change something? The technology is available. The direction is visible. What matters now is execution.

“The Chinese SDV model tends to favour in-house AI development”

As SDVs and AI become central to automotive competition, manufacturers face growing pressure to industrialise both at scale. Dr Liu Qiang, Vice President of Li Auto and General Manager of Li Auto Germany R&D Center, explains how AI, SDVs and embodied intelligence are reshaping vehicle architecture.

Benjamin Müller

Why speed is becoming an industrial requirement

Speed was one of the clearest themes in Ludwigsburg. It is no longer just a competitive advantage. For many automotive companies, it is becoming a basic requirement for survival in global competition.

Sobottka put it bluntly: speed is more important than ever. Harman used the congress to present an approach designed to bring new product solutions much faster from concept to start of production. The underlying issue is broader than one supplier. Carmakers and technology companies must shorten development loops without compromising safety, quality or reliability.

The pressure is particularly visible in relation to China. Several speakers described how Chinese manufacturers shorten development cycles, integrate customer feedback more quickly and bring digital functions to market at high speed. Joachim Langenwalter of TMT CoPilots framed this as a different learning culture. The advantage does not necessarily come from fundamentally different technologies. Many companies use similar tools, chips and AI models. The difference lies in how quickly organisations learn, decide and feed insights back into products.

For European carmakers, that is the challenge. They need to preserve their system competence while becoming much faster at translating test results and real-world usage into concrete product improvements.

How the SDV is turning into an AI-defined vehicle

The software-defined vehicle has shaped the industry debate for several years. At AEK 2026, however, it became clear that the next phase is already taking form: the AI-defined vehicle. Software remains the foundation, but AI increasingly changes how vehicles perceive the environment, prepare decisions, interact with occupants and improve over their lifecycle.

Magnus Östberg, Chief Software Officer at Mercedes-Benz, described this shift in relation to MB.OS and the growing role of AI in the vehicle. One important point was the “jagged frontier” of AI: models can perform impressively in some situations while still requiring strict safeguards in others. For the car, that means AI must not be deployed in isolation. It has to be embedded in architectures that remain verifiable and supported by classical safety mechanisms.

Ned Curic of Stellantis also questioned the term software-defined vehicle from a practical perspective. Customers do not evaluate an architecture concept. They evaluate a car that has to work every day. It must be safe, intuitive, reliable and consistent with the brand. With STLA Brain and STLA SmartCockpit, Stellantis showed how a group-wide platform approach is intended to make functions faster to reuse across brands.

Dr Liu Qiang, Vice President of Li Auto and General Manager of the Li Auto Germany R&D Center, widened the discussion towards Embodied AI. He described the vehicle as an intelligent system that not only processes digital information but also derives physical action from it. Perception, inference and actuation move closer together. Camera and sensor data, large models, chips, operating systems and vehicle dynamics combine into an architecture in which AI increasingly shapes vehicle behaviour.

Passing the baton – Alfred Vollmer succeeds Ricky Hudi

Alongside the technical and strategic debates, the 30th AUTOMOBIL-ELEKTRONIK Kongress was also a personal moment for Ricky Hudi (left), pictured here congratulating his successor Alfred Vollmer. In his opening speech, the long-standing Chairman of the Advisory Board looked back on the development of automotive electronics and of AEK itself, as well as on his own career, which has been closely connected with this evolution.

Hudi described six phases: the early days of individual electronic functions without an overarching architecture, networking via bus systems, the infotainment explosion, the connected car, the SDV and now the transition to the AI-defined vehicle. His review made clear how closely the history of AEK is linked to the technological transformation of the industry. Since his first AEK keynote appearance in 2002, Hudi has remained closely connected with the congress – as a speaker, a member of the Advisory Board and, since 2019, as its Chairman.

Why collaboration is becoming an architecture issue

Many discussions in Ludwigsburg showed that no company can manage the transformation alone. The complexity of software, semiconductors, data, E/E architectures, safety, security and global scalability has become too great. Collaboration is therefore becoming an architectural principle in its own right.

Östberg referred to Bluetooth as an almost historical example of how joint standardisation can create global market success. A similar logic is now needed in automotive if software-defined and AI-defined vehicles are to scale economically.

Concrete examples were visible throughout AEK. Eclipse S-CORE, first presented publicly at AEK 2025, showed how open source is being brought closer to safety-relevant series development. The high number of audience questions around S-CORE underlined how strong the industry’s interest has become in shared software foundations, reuse and qualifiable open-source building blocks.

How BMW is securing its electronics supply chain

Geopolitical tensions, rising costs and growing chip complexity are putting increasing pressure on automotive electronics procurement. At the AEK, BMW’s Florian Weig explained why transparency, standards and new forms of collaboration are becoming critical.

Pascal Nagel

The semiconductor debate pointed in the same direction. Peter Schiefer, President of the Automotive Division at Infineon Technologies, made clear that central computers alone will not be enough. The vehicle still needs microcontrollers, deterministic real-time control, functional safety, energy efficiency and local intelligence. In this context, RISC-V is gaining relevance as an open architecture for future automotive microcontrollers.

Florian Weig, SVP Purchasing and Supplier Network Digital at BMW, focused on the supply chain. In semiconductors, software and AI, traditional linear supplier relationships are increasingly insufficient. Initiatives such as CHASSIS for chiplets, Catena-X and S-CORE stand for new forms of collaboration at deeper levels of value creation.

Why organisation determines execution

Many speakers agreed that the technology is demanding, but organisation often determines whether transformation succeeds. Structures, decision paths, responsibilities and corporate culture are becoming decisive factors in competitiveness.

Langenwalter connected this directly to speed. Companies that want to learn faster must reduce silos and shorten decision paths. It is not enough to make existing processes more efficient. Feedback from development, testing, production and the market must be connected in such a way that genuine learning cycles emerge. Domain experts need to be closer to decisions because they often recognise technical trade-offs earlier than central committees.

Curic described a similar change at Stellantis. The aim is to align software development more closely with clear responsibilities. Teams should spend less time managing internal complexity and more time bringing functions reliably into the vehicle. Modern tools and automation help, but the key question remains who ultimately takes responsibility for the result.

Alwin Bakkenes used Volvo Cars to show how deeply such a transformation affects the way an OEM works. Volvo thinks about software development from the perspective of the vehicle in the field. What happens in real operation flows back into development, validation and updates. That also changes the meaning of start of production. The vehicle is no longer finished at SOP; it enters a phase of continuous improvement.

How the car becomes an orchestrated experience

Another central theme at AEK 2026 was how users actually experience technological change inside the vehicle. Premium quality increasingly emerges where digital services work reliably, operation remains intuitive and the vehicle understands the context of its occupants. Mechanical precision remains important, but it is complemented by a digital experience that reduces friction in everyday use.

Jørgen Behrens, Vice President and General Manager of Google Maps Automotive, showed how maps are evolving beyond classical navigation. The task is no longer just to get from A to B. The system is expected to make journeys more understandable, show the right lane at the right time, integrate charging into daily routines and respond to natural language. With Gemini-based interaction, navigation becomes more of a digital companion that bundles information and reduces cognitive load.

Ivo Muth, Executive Vice President Audi China R&D, explained how strongly customer expectations and digital ecosystems can differ by region. In China in particular, the car is already closely connected with digital platforms, apps and everyday services. For global vehicle architectures, that creates a difficult balance. They must scale worldwide while taking local digital ecosystems seriously.

The semiconductor paradox in today’s cars

Modern vehicles depend on more and more semiconductors, yet that is precisely where the paradox lies: technological progress boosts performance, connectivity and comfort while simultaneously increasing complexity, costs and supply chain vulnerability.

Dr. Oliver Aubel

Sobottka brought the same issue to the level of overall experience. Differentiation increasingly depends on how well individual systems in the vehicle work together. Drivers and passengers do not experience isolated technology domains. They notice whether the interior responds helpfully, whether information appears at the right time and whether AI, sound, display and assistance form a coherent whole. That is where digital vehicle functions become either real added value or digital ballast.

Why semiconductors now sit at the strategic core

AEK 2026 also showed how strongly semiconductors have moved into the strategic core of the automotive industry. For software-defined and AI-defined vehicles, it is not enough simply to add more compute power. What matters is where that compute sits, how data moves, how much energy is required and how safety and security remain protected within the architecture.

Schiefer summarised this with the term “right compute”. Central computers should not take over every task in the vehicle. Some functions belong in zonal controllers, others remain with microcontrollers or intelligent endpoints close to sensors and actuators. The architecture therefore has to be powerful, real-time capable, energy-efficient and economically scalable at the same time.

Dr Ahmad Bahai, Senior Vice President and Chief Technology Officer at Texas Instruments, argued from a similar direction but placed stronger emphasis on distributed intelligence. AI in the vehicle is not decided by TOPS values alone. Data availability, bandwidth, latency and energy consumption are just as important. In many cases, the greatest benefit emerges when intelligence moves closer to where data is created: at the sensor, in an edge node or in a local controller.

The chiplet panel on the first day and the presentation by Aish Dubey, Vice President and General Manager High Performance Computing SoC at Renesas, made clear that new semiconductor concepts cannot be viewed in isolation. Chiplets are more than a packaging option for automotive. They affect the whole system architecture: how functions are partitioned, which building blocks can be reused and how safety, software portability and cost remain manageable.

What remains from Ludwigsburg

AEK 2026 made visible where the pressure to act is greatest. The anniversary left less of an impression as a look back and more as a view of what must happen next. The industry knows its topics, its weak points and its competitors. The energy in the room was clear. Now it has to translate that energy into execution.

The next AUTOMOBIL-ELEKTRONIK Kongress will provide a first answer. On 22 and 23 June 2027, the industry will return to Ludwigsburg. By then, it will become clearer which architectures have matured, which partnerships are working and where the many announcements have already become implementation.

Perhaps the mood will be more festive then. The 30th anniversary was primarily a reality check under pressure. The 31st AEK could become the moment to celebrate first signs of progress.

Live Blog from the AUTOMOBIL-ELEKTRONIK Kongress 2026

After a first day dominated by SDVs, AI and semiconductor strategies, the second day of AEK 2026 in Ludwigsburg continued the discussions around vehicle architectures, automation and the next steps towards series production.

Dr. Martin Large