The Software-Defined Vehicle Has Arrived: Automakers Must Follow Suit

Rajjie Sarmey is Founder and Chief Visionary Officer of FutureProof CXO, advancing governed AI, enterprise strategy and durable value.

gettyFor decades, the automotive industry has talked about the software-defined vehicle (SDV) as though it were primarily an engineering destination: centralized computing, decoupled software and hardware, cloud-connected tools and real-time over-the-air (OTA) updates.

Much of that built-in integration is becoming a reality now. The most complex part of building the software-defined vehicle may no longer be understanding and defining the vehicle. It may be redefining the enterprise itself and the associated ecosystem that now must autonomously and continuously change, secure, govern and improve it.

Today, software touches nearly every aspect and dimension of the vehicle, from battery management to infotainment, braking, cybersecurity and advanced driver assistance. The automotive industry’s movement toward centralized and zonal computing design and architectures is the critical foundation for scalable, open software-defined vehicles and increasingly continuous feature delivery capability.

A vehicle becomes truly software-defined when its capabilities can evolve safely, repeatedly and economically post-production and when the organization behind it can support that evolution without a bottom-up overhaul of its engineering, quality assurance and governance machinery elements every time something happens or changes.

Traditional automotive development was optimized around a familiar progression: engineer the vehicle, validate it, manufacture it and release it. Software altered that premise forever.

When impactful vehicle capabilities continue changing after production, engineering effectively continues for the lifespan of the vehicle. Expectations, software updates, vehicle platforms, vendor-partner interdependencies, cyber resilience, validation evidence, cloud and regulatory obligations must remain harmonized across potentially millions of vehicles operating in different geographies and markets.

Current industry specifications reflect this reality. ISO 24089:2023 establishes specifications for software-update engineering across organizational and project levels and applies across vehicles, systems, ECUs, infrastructure and deployment. Likewise, UN Regulation No. 156 addresses SDV approval concerning software updates and associated software-update management systems.

The modern vehicle is no longer independently built by one enterprise. Its capability is increasingly shaped globally by semiconductor providers, cloud platforms, connectivity companies, AI and autonomy specialists, tier-one suppliers, cybersecurity partners, standards organizations, regulators and regional technology ecosystems.

BMW’s SDV approach provides a practical reference point for this. BMW is collaborating with Momenta on China-specific advanced driver-assistance capabilities, beginning with Neue Klasse, while maintaining its broader systems control strategy of integrating regional technology stacks.

An OEM manufacturer may not need to own every algorithm, platform or component. It does need to control the architecture, integration contracts, security and safety envelope, data rights, life cycle standards and decision authority governing what enters the vehicle and what that software is permitted to do.

We need to address our inclination to fit software into the industrial logic of the traditional automobile. Software is not just another ECU or a feature package. Its value comes from change. That creates friction with organizations pre-wired to freeze configurations, validate extensively and minimize variation before production.

The response is not to wholesale import consumer-software practices, since consequences in the automotive industry can be hazardous. A failed entertainment application is irritating. A defective software change affecting braking, steering, battery management or driver assistance could lead to an entirely different consequence boundary.

Software changes must be accelerated, and so too must governance efforts. Leaders can do this by optimizing manual workflows, automating the mechanics of governance and preserving human accountability for consequential decisions.

Be sure to link all governance requirements to software changes and updates. You can continuously generate test results by enabling machine-readable vehicle configurations. You can also leverage digital twins to conduct further validation. Cyber-resilience checks should operate natively within your engineering pipelines, where software deployments can be staged, observed and reversed.

Human intervention must be a key part of addressing any exceptions. Human workers are not there to simply transport evidence from one system to another.

1. Vehicle architecture covers compute, sensors, actuators, networks and energy systems.

2. Software architecture covers platforms, services, applications and update mechanisms.

3. Enterprise architecture connects engineering, manufacturing, suppliers, data, testing, compliance and life cycle operations.

4. Authority architecture determines who may change what, under which conditions, based on what evidence and with whose accountability.

5. Interconnecting all of the other four is the ecosystem architecture. This includes how the enterprise deliberately shapes its external capabilities, interfaces, standards, data exchanges and commercial dependencies without allowing the vehicle to become an uncontrolled collection of third-party technologies.

I see the operational path to a fully realized SDV in stages rather than around one mythical launch date.

Through the year 2027, architecture convergence will remain predominant, meaning centralized computing, zonal designs, maturing OTA functionality, shareable and reusable software platforms and stronger hardware/software modularity.

Between 2027 and 2029, the heavy lifting shifts to intentional operating-model convergence. And that means integrated engineering pipelines, continuous assurance, automated traceability, cybersecurity-by-design, supplier integration and disciplined ecosystem governance.

Through the latter part of the decade and in the early 2030s, the leading platforms will move toward what I call “life cycle convergence,” where telemetry, digital twins, software delivery, fleet configuration, simulation and operational feedback will come together and increasingly function as one continuous engineering system.

The software and technology required for the SDV to become a reality is taking shape faster than the organizational capability required to operate one is already in place. That organizational capability is, in my view, the real game-changing transformation that the industry needs to address.

The winners of the SDV age must not simply put more software into automobiles. They must reimagine themselves and intentionally architect the ecosystems around them, so that hardware, software, integrators and people can continuously evolve alongside the vehicle without compromising safety, trust or accountability. When that happens, the SDV will become a synchronized operating system for mobility.​​

Forbes Technology Council is an invitation-only community for world-class CIOs, CTOs and technology executives. Do I qualify?