Many experts see the development of SDVs in a transitional phase.Selvi - Adobe.Stock
Few topics occupy the automotive industry as much as the Software-Defined Vehicle. But how far along is the development really? We analyze the various challenges on the way to the car of the future.
A glimpse into the future often requires plenty of imagination. Regarding technological developments, one can usually be even less certain due to the rapid changes. It can help to occasionally look back at the past to better anticipate potential changes: "If you had asked me 15 years ago if I would stop buying physical DVDs, I would have said absolutely not - but now I am happy to subscribe to digital content and stream it on demand," says Jonathan Davenport, Senior Director Analyst at Gartner. With this comparison, the expert suggests that a similar dynamic is likely in the automotive industry. However, it may take several more years before OEMs establish permanently promising monetization strategies for the period after the car sale.
Are current cars already Software-Defined Vehicles?
The question of when a car can be considered a Software-Defined Vehicle divides experts. There is no one hundred percent classification. "We already have vehicles on the road today that deserve this name," says Jörg Tischler, Head of Connected Mobility at T-Systems. Wherever software constitutes a significant portion of vehicle functions, it can be referred to as a Software-Defined Vehicle. Nevertheless, the industry is still at the beginning of development, especially regarding the monetization of features on demand and data. "Customer expectations are changing, and they expect software to always be up to date," says Tischler. This will certainly develop further in the future, but at the moment it is more of a cost factor than a source of revenue.
First, the groundwork must be laid. "I do not expect a quick implementation of SDVs. This will be an iterative process. One of the biggest challenges is the transformation of the electrical/electronic (E/E) vehicle architecture," says Davenport's Gartner colleague Gaurav Gupa, VP Analyst. The automotive industry must transform its software development - especially regarding microservices, containerization, and hypervisors, Gupa continues. Christof Horn, Global Lead Software Defined Vehicle at Accenture, argues in the same direction: "Comparatively new manufacturers like Nio or Tesla have the advantage of being able to set up modern architectures from the start, without having to consider legacy issues like traditional manufacturers."
Here, it is pointless to compulsively try to reach the same level. "Traditional car manufacturers do not have the talent and culture to optimize the development process to shorten time to market," Gupa clarifies. They compete with niche electric vehicle manufacturers from China and established large EV manufacturers like Tesla, which have a much stronger technological base. Therefore, according to Gupa, they should focus more on not making mistakes and achieving smaller goals - moving too quickly leads to errors. Hence, collaboration with hyperscalers is all the more important.
The Opportunity of Open Source
But not only in terms of partnerships for cloud services have tech giants shown the importance they have long held in the automotive industry. "The most successful example of open source at the moment is Android for infotainment - the advantages of a scaling platform have convinced most OEMs," says Christof Horn. To shorten development time, he also emphasizes the importance of collaboration: Open source creates standards in a very efficient way. "Not five years of working groups first, but contribution first: Useful code quickly prevails." Standardization bodies often take longer than technology cycles - that cannot work.
External support is in demand
The Gartner experts have observed a trend that has gone through three phases. Over a period of about four years from 2018, it was observed that automotive manufacturers entered into partnerships with technology companies like Microsoft, Nvidia, and Foxconn. Since 2022, there have been increasing partnerships with IT service companies. "My interpretation of this development is that automotive manufacturers had difficulties in attracting internal development talent and therefore rely on the experience of IT service consultants, while at the same time wanting to own and design the intellectual property that flows into their vehicles," analyzes Davenport. By 2024, a third partnership phase has been reached, in which automotive manufacturers enter into partnerships with each other and particularly with smaller players to gain access to software and electrical/electronic architectures for their vehicles. The latest example of this was the deal between Volkswagen and Rivian.
The direction, away from many control units towards a few powerful domain controllers, is set. However, each manufacturer must define for themselves how quickly and in what technological leaps they implement this, emphasizes Christof Horn. Jörg Tischler also finds that while the trend is clearly recognizable, it should not be seen as a panacea. "Some OEMs are seeking partnerships with large technology companies and using their components, regardless of whether they run in a domain controller or in the cloud," says Tischler. Centralization offers many advantages, particularly in terms of fail-safety and update capability. However, it is important to emphasize that this is not the only approach and the overall architecture must always be considered individually.
New Regulatory Challenges
All hardware and software for vehicles must be certified as automotive-grade. "This is very different from consumer electronics, and the regulations are very strict - no wonder it's a slow-moving industry," says Gupa. While a lot of emphasis was initially placed on features in the development of Software Defined Vehicles, regulatory requirements must now also be met. These have been strongly defined in the last twelve to 18 months, meaning manufacturers need to adjust their processes accordingly, says Jörg Tischler.
When are SDVs financially worthwhile?
While the direction is clear, how quickly the development of software-defined vehicles progresses, as always, depends on money. "It is important to ask what goals one actually wants to achieve with a software-defined architecture. The monetization strategy will be central in the future," emphasizes Horn. The status quo still looks rather poor here. It remains to be seen whether the provision of functions as a service will be successful and whether customers will be willing to change their mindset and acquire vehicle functions in this way. "I am optimistic that this change will come," says Davenport. Which brings us back to the example of DVDs fifteen years ago. But for a similar development to take hold in cars, the shift to Software as a Service must make sense for customers. This, in turn, challenges companies to break out of the usual pattern and be innovative in their pricing.
