Software for the SDV: From the BOM to the Subscription Model

No longer do the outdated metrics that come from the world of mechanics determine how the vehicle faces the customer, and the tried-and-true qualities of vehicle engineering, such as mechanics and manufacturing technology, are no longer enough to remain competitive. Software and cloud platforms are displacing the gaps in the industrial process – and in customers' favor.

David Slump explained in his presentation how supplier Marelli (formerly Magneti Marelli) has adapted to the changing circumstances. In the view of the CEO, competition in the automotive industry in the future will primarily take place in three areas: Powertrain, E/E architecture and software - and the user experience will play the decisive role. Technology, Slump says, has only one purpose: ”It's there to enable the user experience.”

Above all, Slump wants to help his OEM customers get the solutions they want to market faster, and do it all at a lower cost. To that end, he has set his company some content priorities, such as signature exterior lighting and cockpit. In doing so, he wants to provide his customers with holistic support, which means ”we have a focus on hardware AND software,” the Marelli CEO says.

Scaling the vehicle operating system

Kishor Patil's focus is on software. In his presentation, the co-founder and CEO of KPIT Technologies outlined the economic dimension of the transformation to the SDV, or software-defined vehicle: the software development effort required across the industry will amount to approximately $42 billion by 2030, more than double the current value. In his presentation "Scaling the Vehicle:OS - Success factors for the Realisation of Software-Defined Vehicles", Patil shared his company's experience from a software integrator's perspective. According to him, KPIT has realised the importance of the SDV concept for the industry. It leads to a new business model and enables the development of new markets. With the implementation of this concept and the introduction of the software-defined vehicle, Patil expects a massive increase in revenue in this market, which is why the industry will invest around $41 billion in its introduction by 2030.

Patil has identified some key success factors for a successful transformation to SDV. These include a data-first architecture together with an appropriate E/E architecture, open APIs together with scalable middleware and maximum reuse of software assets. To achieve the necessary speed, this must be accompanied by the virtualisation of engineering processes and resources.

Really deep insights into tactical and strategic goals of OEMs and suppliers along the software value chain were provided by a panel discussion under the topic ”New Understanding of Roles in the Eco-System for the software-defined Vehicle” chaired by Ricky Hudi. In the discussion, he explored key aspects affiliated with the transformation towards the SDV. For example, the following question is of significant importance: What are the most important key capabilities for suppliers to remain competitive? Gilles Mabire, CTO of Continental Automotive Technologies, had an answer: ”It depends. It depends on what the specific competence profile of each company looks like.” This determines the added value that a company can contribute. „It is this added value that customers recognize and are willing to pay for,” Mabire said.

Mathias Pillin, division manager of e-mobility at Bosch, advocated a proactive attitude toward technological change: ”We need to embrace change rather than fight it.“ For Pillin, the decoupling of hardware and software – one of the most distinctive prerequisites for SDV – has a good reason: It allows OEMs to achieve the desired customer experience faster and with better results, and also to scale.

A key issue with the SDV is updateability. It must be possible to update the software in the vehicles over the entire life of the car. This not only requires appropriate organizational arrangements at the manufacturer. From the developer's point of view, it means that the installed computer hardware must also meet future requirements. Heiko Schilling, who is responsible for software development at Stellantis, picked up on this point. According to his calculation, the ”technical“ lifetime of a vehicle model – i.e. the development, production and usage phases combined – is theoretically around 15 years. But installing appropriate computers is not realistic, Schilling said: ”We want to develop hardware that can be used for extensive value-added and functional updates in the vehicle over a five-year time horizon. After that period, you can still get security updates.“ An alternative, he said, is to have the option of updating the hardware every quarter over a production period of about eight years.

The issue of the business model for vehicle software emerged as a central point in this discussion. Its volume is increasing at breakneck speed, but its value is inadequately reflected in the industry's current business model – often it has simply been added to the cost of hardware. Jørgen Behrens, head of Google's Geo Automotive, explained why a change in this model is urgently needed: ”People buy cars because they are emotional objects. The iconic character of vehicles is no longer just in their external appearance, but also in the digital experience.“

Against this backdrop, the question of an appropriate method for valuing and monetizing software as a commodity that makes innovation possible and that will increasingly have to be considered independently of hardware in the future arises massively. For Matthias Pillin (Bosch), the ball is in the supplier's court for this discussion: ”It's up to us to quantify this value and make it clear that software is worth its price.“ After all, he pointed out, his company has invested heavily in development and now offers processor-independent software stacks for automated driving – giving OEMs, as his customers, a significant degree of flexibility.

For Heiko Schilling of Stellantis, the question of valuation arises in two directions – firstly in purchasing and secondly in terms of the customer perspective. The currently prevailing valuation scheme between OEM and supplier is the ”BOM model“, in which software is incorporated into the unit price of hardware. But this model is problematic – ”I never understood it,“ Schilling admitted. He added that because software is intellectual property, much like the sales models in the music industry, the valuation of software will evolve similarly in the future – as a subscription business. ”In the future, the sale of a vehicle will just be the starting point for a subscription business – whether it's an ADAS function or infotainment content,“ Schilling stressed. That's very doable, he said; it's also well understood by customers. This change faces mainly internal challenges, he observed: ”The biggest issue for us as an OEM is transforming our purchasing organization into a longer-term subscription-based organization.“