Building Resilience to Automotive Supply Chain Shortages with Software-Defined Vehicles

The automotive supply chain crisis has brought to light the necessity of the software-defined vehicle. The crisis, which traces its origins as far back as March 2020, continues to depress the new vehicle sales market. Successive geopolitical crises, including renewed Chinese lockdowns in early 2022 and the conflict in Ukraine, have extended the semiconductor shortage crisis, which began with Original Equipment Manufacturers (OEMs) cancelling orders in anticipation of an unprecedented demand contraction. Now, the looming energy crisis threatens manufacturing output as temperatures will begin to drop in Europe in a few months’ time. Germany, home to 945 OEMs or automotive component suppliers is particularly exposed, and it is currently expected that a shortage of natural gas would see residential heating prioritized at the expense of industrial output. As the crisis persists, it increases the risk of pent-up demand for new vehicle sales being eroded by high inflation, and interest rates being raised by central banks in response to increasing prices. This will add fuel to the fire that is the current collection of automotive supply chain issues encompassing the industry.

In the face of such challenging macroeconomic headwinds and a protracted supply crisis, automakers are subjecting their long-term, aspirational technology trends to greater scrutiny. As previously alluded to, one such area of long-term automotive investment is the software-defined vehicle, which will see automotive user experiences for both drivers and passengers shaped in software, rather than hardware. Historically, this shift toward software-defined vehicle design has been driven by a desire to deliver dynamic user experiences that can be monetized over the entire lifecycle. However, it is important to recognize that the shift toward software-defined architectures also holds the key to insulating the automotive industry against future supply chain shocks.

The current automotive architecture is hardware-defined, making the automotive supply chain also hardware-defined. Each unique experience that an OEM wants to deliver means a unique hardware iteration of the vehicle platform, whether that is in the form of a separate model or a trim option within a specific model. Each of these Electronic Control Units (ECUs) has its own supply chain, consisting of a Tier One integrator and multiple Tier Two component suppliers. While automotive OEMs have notoriously poor insight into their own automotive supply chain, the task is made almost impossible for a vehicle architecture consisting of some 100 or 150 separate ECUs.

Consolidation of the current, bloated vehicle network into a handful of higher headroom domain controllers or mixed-criticality compute units will enable today’s application-specific ECUs to be defined in software, with many of today’s ECUs hosted on a single domain controller/mixed-criticality processor. Therefore, the driver and passenger user experience of each individual vehicle will be dictated by the software configuration of the vehicle, rather than the hardware configuration, which will be a more common/standardized platform spanning multiple trims, if not multiple models.

This transition will yield multiple advantages for automakers, including the following:

• Better Automotive Supply Chain Insight: For automakers, the task of managing their supply chain has now become a top priority, with maintaining access to critical components now thought to be more important than minimizing inventory. The Herculean task of understanding dependencies in a fragmented supply chain will be made easier, or indeed possible, when the 100+ ECUs of today’s vehicle platforms are replaced by a handful of domain controllers.
• More Generalized Hardware: Defining vehicle experiences in software, rather than hardware, leaves OEMs less exposed to shortages in specific components. An extraneous shock interrupting the supply of a generalized domain controller or application processor could be alleviated by opting for an alternative, generalized processor less impacted by the same shock. Currently, if a vehicle experience is dependent on a hardware component that is in short supply, automakers must choose between opting to ship a vehicle without that experience, or to not ship a vehicle at all. War, natural disasters, and epidemics have far less impact on software development than hardware component manufacturing. Refocusing on a consolidated vehicle platform will enable automakers to focus limited supply chain visibility resources on ensuring a continuous and reliable supply of a handful of domain controllers/application processors, and shaping vehicle experiences in software components, with innovation and production far more resilient than hardware components.
• Alignment with System-on-Chip (SoC) Supplier Investments: A hardware-centric design philosophy has seen automakers increasingly dependent on legacy hardware designs, such as large process node semiconductor designs, to enable a certain experience at the minimum possible cost point. Building a vehicle platform around high headroom application processors built on more modern process nodes will bring automakers into closer alignment with the investment priorities of their SoC suppliers.
• Reduced Weight: Accommodating dozens of ECUs on a Controller Area Network (CAN) bus-dominated, flat network topology has resulted in a complex and heavy wiring harness, which adds weight and cost to today’s vehicles. A more consolidated architecture will enable a simplified and lighter wiring loom.
• Lifecycle Maintenance and Monetization: The original and long-term advantage of the software-defined vehicle (defined by fewer, higher headroom application processors) is the ability to deliver dynamic vehicle experiences, and to monetize connected vehicles after the point of sale. This could be either through one-off purchases or through subscription business models, such as the Ford BlueCruise approach. Fully realizing this revenue opportunity will require automakers to seed the installed base of connected vehicles with excess hardware capacity, with each vehicle featuring more compute resources than needed for vehicle functionality at the point of sale. Currently, automakers pursuing a software-defined vehicle strategy are targeting no more than 50% capacity of an SoC to be required for the vehicle functionality at the point of sale, leaving headroom for future updates throughout the lengthy vehicle lifecycle.

Clearly, some hardware components are non-negotiables when building a vehicle or any other physical device. However, the current hardware-centric approach in the automotive supply chain sees OEMs exposed to shortages of hardware components that could be easily replaced by software components under a platformized, consolidated compute architecture. This pivot would also have the advantage of allowing automakers to focus on ensuring continuous supply of the non-negotiables.

For all the advantages of software-defined vehicle architecture, like safeguarding against automotive supply chain issues, the scale of the challenge facing automakers should not be underestimated. A failure to deliver on the critical software components of key Electric Vehicle (EV) models has seen Volkswagen Chief Executive Officer (CEO) Herbert Diess toppled. When asked about the firing of Diess, Elon Musk, a pioneer in software-defined vehicle architectures responded that software is “the key to the future.” Meanwhile, BMW’s tentative steps into the software-defined vehicle era saw heated seats offered as a subscription option, only to receive considerable media backlash. Automakers and their suppliers are dominated by mechanical engineers, and a consolidated hardware architecture must be met with investment in software development expertise.

BMW’s heated-seat debacle has yielded an important lesson for automakers—that the process of transitioning from a hardware-defined architecture to a software-defined architecture must be carefully managed to maintain consumer acceptance. While taking a feature currently defined in hardware and defining it in software can make a vehicle model more resilient to hardware shortages, automakers should resist the temptation to hide legacy vehicle experiences behind a paywall. General Motors (GM) is another automaker taking a crack at the software-defined vehicle, with plans to launch the Ultifi platform in 2023. The platform will enable drivers to take advantage of subscription-based apps and services that provide greater and more personalized control over vehicle connectivity.

Consumers already accept varied experiences in digitally-native devices. For example, two iterations of a laptop with identical hardware specifications can deliver two completely different experiences for the end users, depending on the software packages and tools that end users install, and often pay for. Therefore, automakers should focus post-sales monetization of software updates on the emergent, digitally-native domains of autonomous driving, connected vehicles, and electrification.