Is the Software Defined Vehicle Already Obsolete?

Rivian’s Chief Software officer recently said that Rivian is moving beyond the SDV era into “AI-defined cars” – what started as a nascent industry term is now officially in an OEM’s lexicon for the first time as far as we know, at least for a non-Chinese OEM. In the meantime, messaging from other OEMs aligns with moving past SDV to an AI-defined future, and suppliers are scrambling to find the use cases that build on AI-based messaging and capabilities.

Does the industry need a new term? Is the distinction so important that it will define future strategies and organizations, a logical evolution of existing SDV programs, or somewhere in the spectrum between?

The basics

Fundamentally, the difference between the two terms is subtle but significant: SDVs are in essence “programmable” vehicles, while AIDVs are “adaptive”.

SDVs enable capabilities like OTA updates, monetization through features on demand (i.e. FaaS), the protection of vital software security patches, and much more. It allows an OEM to add a new feature, improve existing features, and find value in the customer relationship beyond the traditional point of sale and service interactions – it has been a fundamental shift in how we organize ourselves as an industry, and how we develop cars.

The increase in onboard compute power and the rapid rise of AI chatbots-turned-agentic has led to the natural convergence of the two trends into the AIDV, which will enable better predictive diagnostics, actually helpful in-vehicle voice personal assistance (VPAs à AI Assistants), personalized vehicle behaviors & user interactions, better power management, and many more capabilities that will provide value to customers, more revenue channels for OEMs, and more clarity when developing future platforms, vehicles, and products. 

The clearest distinction

SDVs are deterministic; they are data driven, rules-based, and (re)programmable; they represent the digitization of a historically analog product. AIDVs are inferential; they analyze many, sometimes thousands, of data points and signals to reach conclusions that are not strictly following human-defined rules. In some cases, they would also be allowed to make corrective actions.

The clearest use case divergence: diagnostics

Being rule-based, SDVs mainly detect faults, issues warnings & reports, and potentially enable correction through software updates. An AIDV would, in theory, be able to predict future failures, identify root causes, and recommend corrective actions while automating service workflows in the background. 

The autonomous driving question

The most fundamental impact of the surge in AI capabilities over the last few years has been the resurgent viability of autonomous driving capabilities. While SDVs offer OEMs the ability to introduce new capabilities, refine and maintain existing ones, and introduce new efficiencies and improvements, AIDVs can continuously improve fleet-wide driving behavior or for the specific vehicle to align with user preferences and habits.

Variations in the tech stack

From a technical perspective, the changes to the technology stack that enables AIDVs is an evolution of the SDV tech stack, although at least some internal reorganization will be needed at OEMs intending to go the AI path, especially on the R&D side.

결론

While the first instinct might be to abandon SDVs for the more attractive, more innovative-sounding, and more sophisticated AIDV, it is important to consider that the SDV remains the foundational enabler of the AI capabilities that will power the AIDV.

The good news is that massive reorganization will not be needed, at least in the short term, to accommodate AIDVs. The bad news is that this is the result of an imbalance in the industry: while some OEMs are close to having truly software-defined vehicles, most of the others are still managing this fundamental transition. 

“If anything, this only proves that getting SDV programs and initiatives right is essential to the future feature roadmap; not only does it enable short term introduction of capabilities, iterative improvements, and new business models, but it the fundamental piece for the future of vehicle development and feature deployment. The AIDV will be built on the SDV, without this wasn’t the initial roadmap for SDVs. That is the true power of software enablement."  Mo Al-Bodour, SBD Automotive Senior Manager

How SBD can help

SBD Automotive can help benchmark your position against the wider industry and identify where action is needed most. To explore how these trends impact your strategy, architecture and supplier roadmap, get in touch with SBD Automotive for a deeper discussion. Email info@sbdautomotive.com