Vehicle software and software-defined vehicles

The digital transformation of the car industry is most evident in the emergence of software-defined vehicles (SDVs), in which software determines an increasing share of vehicle functionality. This shift is not based on software alone, but also on a re-imagining of the electronic and electrical architecture of vehicles over the past decade, pioneered by pure-play EV makers. In conventional distributed architectures, each function – such as lighting, braking or climate control – is managed by its own dedicated controller, or electronic control unit (ECU). This is now giving way to domain or “zonal” architectures, with a smaller number of ECUs controlled by central computers.

This shift reduces wiring complexity and enables a greater share of vehicle functionality to be defined and updated through software, and expanded over time1. As a result, critical functions such as advanced driver assistance systems (ADAS) and battery management systems can be improved through over‑the‑air (OTA) updates. Manufacturers can also deploy new features, performance improvements and security updates more rapidly across a vehicle’s lifetime. When combined with the lower material requirements of zonal architectures compared with distributed designs, these benefits can reduce EV production costs, provided volumes are sufficient to spread development costs across enough vehicles.

Lower vehicle production costs can translate into lower purchase prices for consumers, but an increasing number of features in base models and greater reliance on subscription‑based functions could also increase costs for users, albeit creating a new revenue stream for manufacturers. For example, “feature‑as‑a‑service” business models – through which users can access premium vehicle functions through one‑off payments, subscriptions or pay‑per‑use arrangements – offer greater flexibility but may also raise lifetime costs depending on automaker strategies and consumer choices.

At the same time, software – particularly over‑the‑air updates – can support affordability by ensuring the vehicle remains up to date for longer. First introduced by Tesla in 2012, OTA updates enable software and firmware upgrades delivered via wireless connectivity, without requiring a visit to a dealership or service centre. OTA capabilities enable manufacturers to fix software defects, improve vehicle performance, tune vehicle functionality, deploy cybersecurity patches and introduce new features after the initial sale.

While most major automakers have plans to develop these software-defined systems across different powertrains, all currently available models with zonal architectures and extensive OTA capabilities are battery electric – primarily from pure‑play EV manufacturers. Additionally, while conventional vehicles and SDVs represent two ends of a technological spectrum, many automakers are currently deploying architectures that fall between them. These intermediate systems consolidate some functions into shared computing units but retain dedicated controllers for safety‑critical or legacy systems. This gradual approach allows manufacturers to expand software capabilities and enable additional features at the same time as managing development costs and timelines associated with the required vehicle and production platform redesigns.

Software development requires a fundamentally different approach to traditional automotive engineering. In conventional vehicle development, specifications are typically defined early in the design process. In contrast, software development relies on iterative cycles in which a set of features is introduced early on, followed by multiple rounds of updates for bug fixes and additional functionality. This shift can create advantages for new entrants structured around a software-first model and for incumbent manufacturers that succeed in adapting their processes, but it poses a challenge to companies that do not.

The role of electronics in cars has been growing ever since the 1970s, steadily expanding vehicle functionalities. A significant turning point occurred in 2017, when Tesla shifted from a distributed electrical and electronic architecture to a centralised, software‑defined approach.

However, successfully transitioning from a historically hardware‑centred business model to software‑defined vehicles is neither quick nor cheap. In 2023, Volkswagen scaled back its ambition to develop core vehicle software entirely in‑house though its own software company, CARIAD. Instead, it shifted towards partnership‑based models, including the establishment of a joint venture with Rivian. Similarly, Ford decided to abandon its “fully networked vehicle” project in 2025. Other incumbent automakers are, however, shifting more smoothly towards software-centred vehicles, such as BMW and Mercedes-Benz.

The increasing software content of vehicles is also reshaping the role of operating systems. A growing number of automakers are adopting automotive-grade operating systems derived from consumer technology ecosystems, such as Android Automotive OS, to support infotainment, navigation and app-based services. This trend can lower development costs and accelerate feature deployment, but it raises strategic questions around dependency on large technology providers, data governance and long-term control over the in-vehicle user interface.

Automakers are adopting a range of strategies to strengthen their software capabilities. Companies such as Tesla, Toyota, and BYD emphasise in-house development to retain control over key vehicle functions and data. Others have established dedicated partnerships, such as General Motors and NVIDIA, or Renault and Google, software subsidiaries like BMW Car IT, or an in-house software platform such as General Motors Ultifi.