As the automotive industry accelerates toward the Software-Defined Vehicle (SDV), the conversation is rapidly shifting from ambition to execution. In this interview, Miroslaw Staron, Professor at Chalmers University of Technology and the University of Gothenburg, shares his perspective on the architectural transformations required to make SDVs a reality. As moderator of Track A – Electronics & Architecture at VECS 2026, he highlights a fundamental shift: moving beyond hardware-centric thinking toward truly decoupled, software-first architectures.
Staron explores the industry’s current misalignment between powerful hardware investments and the lack of robust middleware, the delicate balance between centralized and zonal architectures, and the need for hybrid models that enable both flexibility and safety. He also emphasizes the critical role of collaboration in building scalable ecosystems and outlines what will ultimately define successful SDV platforms in the years ahead – adaptability, resilience, and continuous value for the end user.
As the moderator of Track A, what do you see as the most critical architectural shift needed to truly enable the Software-Defined Vehicle?
We often use the cliché ‘smartphone on wheels,’ but a true SDV is far more complex. The most critical shift isn’t just adding more code; it’s the radical abstraction of hardware. To truly enable the SDV, we must move away from the current landscape of fragmented ‘Unix-clones’ on wheels toward a unified, reusable architecture. We need APIs that allow user functions to be portable across models and generations. If the software isn’t decoupled from the silicon, it isn’t an SDV –it’s just a traditional car with a bigger screen.
The track highlights a gap between SDV ambitions and current E/E architectures. Where do you think the biggest misalignment lies today?
The misalignment is philosophical as much as it is technical. We are installing High-Performance Computers (HPCs), yet we are still thinking like traditional OEMs. We’re building powerful boxes without the necessary abstraction layers that guarantee portability, resilience, and robustness. We have the ‘muscle’ (hardware) but haven’t yet developed the ‘nervous system’ (middleware) required to manage it effectively.
We’re hearing a lot about turning SDV investments into new revenue streams. From your perspective, what needs to happen architecturally to make that transition viable?
To unlock new revenue, the vehicle must evolve from a product into a platform. This requires a ‘middle way’ in architecture, but perhaps not a fully fledged ‘middleware’. Traditional layered architectures are too rigid and closed, while pure Service-Oriented Architectures (SOA) can struggle with the latency and safety requirements of a vehicle. The winners will implement a hybrid model – one that offers the flexibility of open APIs for third-party integration while maintaining strict runtime monitoring and traceability for safety and compliance.
There’s a strong push toward centralized and zonal architectures – what are the biggest trade-offs OEMs need to be aware of when making this shift?
The shift to centralized and zonal E/E architectures is a double-edged sword. The biggest trade-off is emergent complexity. In a centralized HPC, you face a ‘tragedy of the commons’ – multiple software components competing for CPU, memory, and bandwidth. If not managed, their cumulative demand can exceed 100% of the system’s capacity. Meanwhile, zonal architectures introduce new challenges in communication orchestration. We are still mastering the art of balancing inter-HPC speed with intra-HPC reliability.
Several sessions touch on open platforms and service ecosystems. How important is industry collaboration in shaping the future SDV architecture?
Collaboration is ‘Priority One.’ As an industry, we have to stop competing on plumbing and start competing on experience. We need to standardize the underlying platform so we can focus on value-adding software. For the consumer, the goal is ‘evergreen’ value. People want a vehicle that feels fresher and more capable five years after they bought it. Collaboration is the only way to build an ecosystem where software can actually scale.
If we fast-forward five years, what do you think will define a “successful” SDV architecture – and which players are best positioned to get there?
Five years from now, a successful SDV architecture will be defined by its robustness and adaptability. It must be impossible to ‘brick’ the car with a software update, and the platform must be open enough for third-party innovation.
But the real hallmark of success will be runtime reconfiguration. Imagine buying a car in Spain where the architecture prioritizes thermal management for AC, then moving to Sweden and having the vehicle’s profile seamlessly pivot to optimize cabin or steering-wheel heating and battery pre-conditioning. The player who makes the vehicle adapt to the user’s changing life – rather than forcing the user to buy a new car – will win the market.[/vc_column_text][/vc_column][/vc_row]