Software Defined Vehicles – the hype and the hurdles

A new technology or evolving trend, according to Gartner, often goes through a hype cycle, which begins with great expectations. This hype is typically followed by a phase of ‘disillusionment.’ Eventually, key challenges are addressed, leading to wide adoption. The software defined vehicle (SDV) is arguably in the initial phase of this cycle, as it confronts the challenges of execution. In a recent panel at the SDV 2024 USA event in Ann Arbor, MI, which I moderated, several COVESA members explored the current state of SDV adoption and the hurdles that must be overcome for SDV to mature.

The panel titled “SDV: The Path Forward and the Challenges Along the Way” included representatives Daniel Cashen and Pete Brown (from Stellantis and Karma Automotive), Brian Carlson from NXP, and Greg Morningstar from Bosch – a nice mix of OEM, semiconductor, and Tier 1s. The discussion covered various facets, including the role of consortiums like COVESA, the impact of SDV on the automotive value chain, the issue of software ownership by OEMs (i.e., make vs. buy), and the influence of new initiatives like the SDVerse marketplace on the software-centric transformation of the industry.

Role of COVESA and Other Consortiums

There was unanimous agreement on the importance of standardized data formats and APIs, as well as the crucial role played by organizations like COVESA. This standardization is foundational to the broader philosophy of reducing redundancy, so OEMs, Tier 1s, and semiconductor companies do not have to reinvent the wheel. Instead, OEMs can focus their resources on developing differentiating features and functions. The COVESA Vehicle Signal Specification (VSS) is already providing value to the ecosystem by enabling multiple stakeholders to communicate using the same language across in-vehicle, mobile, and cloud applications.

The importance of umbrella organizations such as the SDV Alliance (often referred to as a “consortium of consortiums”) was also recognized for ensuring that the efforts of various consortiums are aligned. It was clear that recent initiatives, such as the SDVerse marketplace, are complementary rather than contradictory to the work of these consortiums. SDVerse will allow automotive software buyers and sellers to easily connect, providing a centralized platform for software procurement.

Shifting Supply Chain & Make vs. Buy

While the roles of OEMs and Tier 1 suppliers may be shifting, Tier 1s have the opportunity to form new partnerships in this evolving paradigm, where the traditional tiered supply chain hierarchy is changing. As for the evolving role of semiconductor companies, we learned that they, too, are internally developing their own software and systems expertise. They are increasingly partnering with OEMs, Tier 1s, and software providers, offering reference designs and pre-integrated software environments to address integration challenges.

Regarding the “make vs. buy” (or build vs. buy) debate, OEM representatives on the panel emphasized that software ownership should be focused on what differentiates the OEM — typically applications — rather than underlying enablers like device drivers or operating systems.

The Role of Architecture – Clean Sheet vs. Evolutionary

The advantages of new energy vehicle (NEV) OEMs (e.g., Xpeng, Rivian) in adopting clean sheet architecture, compared to the relative inertia of traditional large OEMs, were acknowledged by the panelists. However, the large OEMs also have clear potential long-term advantages, as the eventual scale of software reuse will be significant. In this context, designing the architecture correctly is critical. The more well thought out and scalable the architecture, the more cost-efficient it will be to adapt to changing needs and support the diverse product portfolios of large OEMs. In other words, architecture matters!