The Rise of the Software-Defined Vehicle

The automotive industry is undergoing multiple profound transformations in parallel. For decades, innovation has largely focused on mechanical engineering. Today, software is becoming a core driver of value creation. The emergence of the software-defined vehicle (SDV) is reshaping how cars are designed, built, sold, updated, and experienced throughout their lifecycle.

 

This shift goes far beyond adding more code to vehicles. It requires fundamental changes in electrical / electronic (E/E) architectures, development processes for both software and the overall vehicle, as well as in supplier relationships and business models. 

 

 

Decoupling Software from Hardware Development Cycles

Traditionally, automotive software development was tightly coupled to vehicle hardware cycles. Major software updates coincided with new model launches or hardware refreshes. Suppliers delivered embedded software integrated into dedicated electronic control units (ECUs), and once deployed, systems were largely fixed.

 

The SDV paradigm breaks this dependency. Pioneered at scale by Tesla in 2012, this model decouples software from vehicle hardware development cycles. Software becomes modular, reusable across platforms, and designed for continuous deployment. Abstraction layers between hardware and software enable updates without requiring physical redesigns. This approach reduces development time and costs while enabling continuous improvement. Vehicles can receive new features, performance enhancements, and bug fixes long after delivery. 

 

 

From Distributed ECUs to High-Performance Computing

Conventional vehicles rely on dozens — over 100 in some cases — distributed ECUs, each responsible for a specific function such as braking, infotainment, or HVAC. This architecture struggles with scalability, data throughput, and upgradeability.

 

SDV replaces this patchwork with a simplified architecture built around high-performance computers (HPCs) and controllers. The first step is often a domain-based architecture (such as an HPC for all chassis-related functions) and ultimately centralized compute with zonal controllers. Rivian, for example, reduced the number of controllers in its R1 platform from 17 to 7 which simplified wiring, resulting in 2.6 km less cabling. BMW’s Neue Klasse architecture relies on four domain-specific “superbrains” connected to four zonal controllers. The latter essentially perform data triage and power distribution functions within a physical area of the vehicle whereas computing is concentrated in HPCs.

 

These new architectures require upgraded on-board networking solutions with higher bandwidth. In addition, the deployment of ADAS functions generates increased data flows and demands low latency to maximize safety. To this end, ethernet is gaining traction, in some cases with optical cables, to connect HPCs with zonal controllers.

 

 

Reinventing the Supply Chain

The SDV transition disrupts traditional supplier models as well. Instead of buying ECUs with embedded software along with the hardware they control, OEMs are investing in modular and abstracted software stacks, scalable computing platforms, and increasingly open ecosystems. Major suppliers such as Bosch, Magna, Denso, ZF, and Valeo have expanded into HPCs and software integration. Those that lack electronics expertise or the capability to develop HPCs or controllers will simply lose ECU revenue streams. At the same time, some OEMs such as Tesla, BYD, Xpeng, and Rivian develop their own software stacks, HPCs, and even AI chips.

 

Collaboration with tech giants, semiconductor designers, and cloud providers has become essential. Companies like Nvidia and Qualcomm, formerly Tier 2 or 3 suppliers (if at all suppliers), now engage early on in development processes with OEMs. Industry-wide initiatives aim to harmonize standards and develop open-source solution, recognizing that interoperability is critical. 

 

 

Over-the-Air Updates and Lifecycle Value

One of the most visible benefits of SDV is over-the-air (OTA) software updates. Carmakers can push feature enhancements and software-related recalls without dealership visits, thus reducing friction for consumers. Tesla demonstrated the power of this capability by remotely improving braking performance within days of receiving negative feedback when Model 3 was launched. Likewise, Rivian introduced or updated hundreds of features OTA in its first year of operation. This capability improves customer satisfaction and should protect residual values.

 

OTA updates also create new monetization opportunities as features can be unlocked after purchase via subscriptions or one-time payments. For example, Tesla offers quicker accelerations for a fee, Mercedes sells horsepower boosts via subscription, and Ford offers BlueCruise, a Level 2 ADAS, for a monthly fee. OEMs have been ambitious about potential recurring revenue. Back in 2021 Stellantis and GM each projected software-related revenue in the tune of 20 billion dollars by 2030. However, monetization has proven more complex than expected. Charging extra for previously standard features such as heated seats triggered backlash. Instead, success will depend on delivering genuine additional value such as a WiFi hotspot or driving assistance upgrades.

 

 

A Challenging Transition 

The shift to SDV is not an evolution but rather a deep, multi-faceted transformation. Some of the challenges for incumbents are technical. Hardware-software compatibility as well as cybersecurity will need to be sustained over a decade or more after any vehicle is sold. Connectivity standards evolve, e.g., 5G to 6G, requiring forward-looking hardware design. Carmakers must also ensure sufficient computing headroom to support future software stacks without costly hardware retrofits.

 

Challenges exist also at corporate level, including culture and talent gaps, engineering processes, and supply chain management. VW’s software entity CARIAD experienced repeated delays in their development, causing vehicles to be launched late. Consequently, future VW Group vehicles will use E/E architectures and software stacks jointly developed with Xpeng for China and Rivian for the rest of the world. Similarly, in mid-2025 Ford cancelled its FNV4 (Fully Networked Vehicle) project, opting for the less ambitious upgrade of the existing architecture.

 

Earlier this year, McKinsey predicted that 47 percent of all new vehicles sold globally in 2030 would still use a distributed architecture whereas 41 percent will feature a domain-based one, and 12 percent a zonal architecture with centralized compute. The latter share is expected to reach “only” 33 percent by 2035. Still a long way to go.

 

 

Leaders at the Digital Frontier

Digital-native players are advancing fastest. According to Gartner’s Digital Automaker Index 2025, leaders include Tesla, Nio, Xiaomi, Xpeng, Li Auto, and Rivian. Xiaomi’s very first cars illustrates the convergence of consumer electronics and mobility. The SU7 and YU7 integrate seamlessly with smartphones and home ecosystems, running proprietary operating systems on a centralized architecture. Incumbent OEMs have no choice but to undergo this transition. According to a 2025 survey performed by McKinsey, 38 percent of premium car owners in Germany would consider switching brands if the alternative offered a better digital experience. 

 

 

Towards the AI-Defined Vehicle

While SDVs are still being deployed at scale, the next stage is already emerging as seen at this year’s CES: the AI-defined vehicle (AIDV). Building on centralized computing and rich connectivity, AIDVs will integrate advanced AI models capable of learning, reasoning, and adapting. Instead of deterministic software controlling predefined behaviors, multimodal AI systems will continuously improve through cloud and vehicle-based data flywheels. AI agents will manage not only in-car assistants but also lifecycle functions from manufacturing to predictive maintenance. The vehicle will evolve from being updatable to being adaptive.