How automotive computing will shape the next vehicle generation

The software-defined vehicle has moved steadily towards the centre of automotive strategy in recent years – including at the AUTOMOBIL-ELEKTRONIK Kongress. Yet the more software, AI and digital functions dominate the discussion, the clearer the limitation becomes: without the right computing and electronics architecture, the SDV remains a promise. As long-time AEK contributor Heinz Leiber put it: “Software matters – but the car still has to drive.” In practical terms, automotive computing helps determine which functions are possible, how safely they operate, how much energy they consume and how long an architecture can continue to scale.

“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

At AEK 2026 in Ludwigsburg, the discussion therefore extended well beyond software platforms and AI models. SoCs, microcontrollers, chiplets, edge AI, Ethernet and RISC-V repeatedly moved into focus. The lesson from Ludwigsburg was clear: the future vehicle will not be decided by code alone. It will depend just as much on the computing foundations that can carry digital and AI-based mobility into the real world.

Why is the SoC becoming the centre of the architecture?

One of the event’s most memorable lines came from Peter Schiefer, President and CEO of the Automotive Division at Infineon Technologies. With “We rock around the SoC”, he described the growing importance of central systems-on-chip in future vehicle architectures. Vehicles are increasingly being organised around powerful compute nodes that consolidate functions, process data and support software platforms.

At the same time, Schiefer made clear that the objective is not to concentrate as much processing power as possible in as few locations as possible. His central point was: “The future of the SDV is not ‘all compute’, but the ‘right compute’.” In other words, the vehicle needs the right level of processing in the right place. Some tasks belong in central high-performance computers, while others should remain close to sensors, actuators or safety-critical control units.

That is why the microcontroller remains a core element of automotive computing. Central computers can assume more functions, but functional safety, real-time behaviour and redundancy cannot be addressed by large SoCs alone. Safety-critical functions still require robust, distributed and energy-efficient control. The architecture of the future will therefore combine centralised performance with distributed safety and control.

AEK 2026: Why execution now matters more than strategy

The automotive industry has defined its technological direction. Now comes the harder part: delivering at scale. AEK 2026 showed why SDVs, AI, chiplets and China-Speed are turning execution into the decisive competitive factor.

Dr. Martin Large

Why are chiplets more than advanced packaging?

The hardware focus became particularly visible during the chiplet panel discussion. Chiplets are seen as one possible way to make the growing complexity of future vehicle computers more manageable. Monolithic high-end SoCs are becoming more expensive, more difficult and riskier to develop, while requirements for AI acceleration, memory access, safety, security and energy efficiency continue to rise.

Chiplets promise a more modular route. Rather than integrating every function on one large die, specialised building blocks can be combined within one package. That is especially relevant to automotive applications, where vehicle platforms remain in service for long periods but still have to accommodate rising performance requirements.

Aish Dubey, Vice President and General Manager High Performance Computing SoC at Renesas, argued in Ludwigsburg that chiplets should not be treated as a packaging question alone. The real issue is architectural: which functions should be separated, which interfaces are required, how can individual blocks be reused and how can the complete system remain manageable over many years?

The panel also made clear that chiplets will only work if technology and business models align. Connecting individual dies is not enough. The industry needs standards, robust interfaces and viable ecosystems, together with demonstrable safety concepts and meaningful cooperation between OEMs, semiconductor manufacturers, suppliers and integrators. Initiatives such as CHASSIS are intended to address precisely this gap.

How is edge AI redistributing compute?

Dr Ahmad Bahai, Senior Vice President and Chief Technology Officer at Texas Instruments, placed edge AI at the centre of his contribution. His perspective addressed a fundamental computing question: where should intelligence reside in the vehicle if it is to operate quickly, efficiently and safely?

Future vehicles will not collect every piece of data and then send it to one central processor. For sensors, driver assistance and safety-related functions in particular, it can be useful to process data close to the sensor or at an edge node. Local preprocessing reduces latency and network load while allowing the system to respond more quickly.

Bahai described the vehicle as an intelligent system that perceives its surroundings, processes data and reacts in real time. This changes how automotive computing should be assessed. The TOPS figures of central processors are only one part of the equation. Analogue front ends, sensor interfaces, embedded processing, power management and an architecture that organises data movement efficiently are equally important.

That matters because moving large volumes of data consumes bandwidth and energy. As more sensors, cameras, radars and AI functions enter the vehicle, the industry must decide which information genuinely needs central processing. Edge AI is therefore not an additional topic at the margins of vehicle computing. It is becoming a core element of future architectures.

Save the date: 31st AUTOMOBIL-ELEKTRONIK Kongress

The 31st International AUTOMOBIL-ELEKTRONIK Kongress (AEK) will take place on 22 and 23 June 2027. For many years, the networking conference has brought together leading decision-makers from the automotive electronics sector and senior executives from the technology industry to discuss the integrated customer experience required for the vehicles of the future.

Despite its increasingly international profile, participants still describe the AUTOMOBIL-ELEKTRONIK Kongress as an “automotive family gathering”.

Secure your conference ticket for the 31st AUTOMOBIL-ELEKTRONIK Kongress in 2027. You can also follow the AEK LinkedIn channel and #AEK_live.

Why does automotive computing extend beyond processors?

Schiefer also stressed that automotive computing does not end with the processor. As vehicles centralise more workloads while moving other intelligence towards the edge, communication, control and standardisation come under greater pressure.

Ethernet is becoming the backbone of future in-vehicle networks. Central computers, zonal controllers, sensors and intelligent endpoints must exchange high-bandwidth data reliably. The more heavily a vehicle depends on software and AI functions, the more important a scalable network becomes for carrying these data flows.

Capgemini Puts AI in Automotive Engineering to the Test

Capgemini’s Automotive Engineering and R&D Pulse 2026 shows how far AI has moved beyond pilots. Executives report shorter concept phases, rising productivity gains, higher budget shares and pressure to reorganise engineering.

Ronja Schmiedchen

At the same time, RISC-V is gaining relevance as an open instruction-set architecture. In microcontrollers in particular, it can support more flexible platforms and longer-term scalability. Open standards are attractive to the automotive industry because they may reduce technological dependencies and give companies greater control over their own roadmaps. RISC-V is consequently becoming part of the wider debate around technological sovereignty and scale.

What determines how far AI can go in the car?

AEK 2026 showed that automotive computing is becoming a strategic core issue. Defining software functions and AI applications is not enough. The industry must also decide which hardware will run them, how data will be distributed, how much energy is available and how safety can be secured across the full architecture.

Several conclusions follow. Processing power alone is insufficient; the decisive factor is an architecture that combines central performance, local intelligence and safety-critical real-time control. Chiplets could help contain the complexity and cost of future vehicle computers, but only with standards and sustainable ecosystems.

Edge AI moves intelligence closer to sensors and actuators where latency, energy consumption or data volumes require it. Ethernet and RISC-V are not side issues, but parts of the same computing discussion. Together, these elements show why hardware sets the pace for what software and AI can deliver in the vehicle.

Anyone seeking to scale new functions must therefore consider the physical platform early enough – from SoCs and chiplets to processing at the edge and in-vehicle networking.

Sponsored

Europe Automotive Computing Conference

Munich on November 18-19, 2026

Why does the Automotive Computing Conference pick up here?

AEK 2026 made the broad direction visible, but many questions are only beginning to reach industrial implementation. Which chiplet roadmaps are realistic? How can safety be demonstrated in modular semiconductor architectures? Which tasks belong in central computers, which in zonal controllers and which directly at the sensor? And how can computing performance create customer value that continues to develop over many years?

These questions will be explored further at the Automotive Computing Conference on 18 and 19 November in Munich. The event will examine the technical foundations, roadmaps and architecture decisions behind the vehicle of tomorrow. AEK showed why automotive computing is becoming a strategic core issue; the conference will continue that discussion with a focus on hardware, semiconductors, compute architectures and practical implementation in the vehicle.