How open source software can speed the transition toward the software-defined vehicle

Munich, June 2023

How open source software can speed the transition toward the software-defined vehicle

Munich, June 2023

oftware development and deployment is emerging as one of the biggest challenges to incumbent OEMs’ dominance of the global automotive market. Established automakers need a new software innovation model, and moving to open source software is a compelling strategy.

When Chinese automakers lifted the curtain on their latest state-of-the-art digital vehicle innovations at the 2023 Shanghai Motor Show they laid down a marker for all automakers. A new breed of OEMs had arrived, and they had digital in their DNA. By building and deploying auto software faster and more efficiently than their incumbent rivals they proved that software is already one of the biggest challenges to incumbent OEMs’ dominance of the global automotive market.

Why is this, when incumbent OEMs have decades of experience in refining auto technology development cycles? Our experience with multiple established brand automakers suggests that standard innovation and supplier management practices are failing to meet the demands of rapid, iterative and collaborative software evolution. These ways of working are unable to cope with the number of software variants demanded, they result in conflict with legacy systems, they fail to deliver effective abstraction layers for simplification and to generate sufficient reusable content, and they are too slow.

Underlying these challenges is the fact is that the standard innovation model does not encourage equal-partner collaboration. Creating the software-defined vehicle demands digital-first skills, and integrated working models that reflect the collaborative ecosystem in which software development flourishes.

Incumbent OEMs have already undertaken several approaches in their attempts to gain momentum in the rapidly evolving field of automotive software. They have tried strategic alliances with big tech suppliers of proprietary software. They have tried building their own dedicated software houses. They have tried reproducing the agile working models familiar to IT businesses, grafting some elements of these onto legacy methods of planning and executing. None of these have really worked – and meanwhile, development costs are ballooning and returns on investment in data are disappointing.

It is time for a new strategy.

Why OSS?

Many established automakers gravitate naturally to the proprietary or ‘closed source’ software model. This instinct is in their manufacturing DNA, having often spent decades refining their own brand-defining proprietary technology. Patented software offered by technology companies such as Microsoft and Adobe comes with the advantage of IT development and integration support – but users are unable to modify or add to the source code. Open source software (OSS) is different: the source code is freely available for all users to inspect, improve, edit or otherwise develop, and under most OSS licenses the only limiting requirement is for the user to publish their code modifications in the software repository.

The OSS model is tested and widespread. Although users may not realize it, more than 70% of all software in use is open source. Developers and end customers often prefer OSS for its superior stability, security and capacity for rapid development and bug elimination; this is one reason why in 2019 IBM acquired Red Hat, one of the world’s largest OSS developers. In many respects, the innovation model behind OSS resembles the unstructured collaborative model that many automakers are already trying to emulate. So it is surprising that, with the exception of infotainment, where OSS operating systems such as Android are already in use, software under OSS licenses still plays a very limited role in the automotive sector.

The OSS approach can be shown to lower cost and improve innovation speed, and thereby increase competitiveness. It reduces cost because of the resource sharing that is intrinsic to OSS, and also minimizes contractual complexity through community ownership of software developments. It lessens technical complexity with better documentation through the simplification of data (usually known as abstraction) that is part of the OSS model, and it helps users to grow collaborative learning and development skills to a level that is unlikely to be achieved in proprietary software provider/user relationships.

Some automakers already understand these advantages and are deploying OSS at scale. One example is Mercedes, which has for six years now been employing an open source strategy that combines free and open source software. Auto supplier Bosch has co-created an OSS-focused ‘’ initiative to improve adoption of digital best-practice in the automotive industry and to generate a repository of interoperable software development tools. Bosch is also involved in the ELISA (Enabling Linux In Safety Applications) project, an initiative from the OSS operating system Linux designed to develop and win certification for Linux-powered safety-critical applications with wide ranging uses in the automotive industry. And most recently, Bosch subsidiary ETAS announced a strategic collaboration with aforementioned Ret Hat on in-vehicle basic software layers. A great range of ‘mixed-criticality’ OSS applications is also being developed by the SOAFEE (Scalable Open Architecture For Embedded Edge) project supported by VW and auto supplier Continental.

A compelling risk/return equation

Many organizations see OSS as a risk to their business, and to some extent their concerns are valid. Yet the rewards for OSS adoption in an environment where companies need to shrink technology development timelines and embrace more collaborative ways of working are great, while the potential pitfalls are over-estimated. The challenge for companies is to minimize risk, and leverage the inherent advantages of an open source approach.

Companies are often unwilling to invest resources in a technology base that they do not own – but proprietary software users do not own the software they deploy either, they merely rent it. They also have questions over security, safety and regulation of OSS applications and operating systems – but addressing these issues directly through the growing number of OSS deployment initiatives may be a better strategy than outsourcing responsibility for mission-critical aspects of their technology stack.

To diminish risk, companies should consider adopting OSS as a core approach where certain conditions are in place. The levels of standardization and abstraction in the selected OSS should already be high, to control development costs and maximize usability. The potential resource savings from OSS adoption should also be high. And finally, the relevance of differentiation should be minimal, meaning that OSS adoptions should concentrate on the lower levels of the software stack, rather than on customer-facing applications where software becomes visible and differentiation is commercially important.

The path to adoption

Automakers are not short of choices for OSS development platforms. They include the Automotive Grade Linux Project, a Linux Foundation collaborative project to bring together suppliers and automakers and create a complete software stack for the connected car; Kubernetes, a software ‘container’ operating sub-system originally designed by Google engineers that is emerging as a powerful cloud application for managing multiple data hungry objects such as cars; and the Red Hat In-Vehicle Operating System, which is extending Linux to driver assistance systems.

As with all innovation choices, opportunities must be matched with real world organizational needs and capacities. Automakers will need to develop a detailed picture of where and how OSS solutions can be applied in their software stacks, and to encourage and enable suppliers to adopt OSS approaches, which include providing resources for joint projects and legal liability cover.

The supplier base cannot be fully leveraged without accepting that a degree of risk-taking is inherent in the collaborative model. Since competitive pressure from digitally native automakers will eventually enforce that model sooner or later, it makes sense for established OEMs to adopt it now and begin to benefit from a reduction in the development burden. The opportunity is there to take action before the industry’s widely-recognized strategic crisis turns into a financial one.

A narrowing window

Today, there is a growing risk that automakers begin to lose innovation capacity to both established technology companies and start-ups, as skilled employees see that the attempt to create software-defined vehicles on the old siloed and proprietary development model is failing. Yet OEMs still have know-how and structural advantages that they can leverage: the automotive industry remains dependent on domain-specific expertise and the ability to orchestrate technology and suppliers in a complex regulated environment.

Established automakers understand issues of feasibility and the real value-add of technology. They appreciate that new approaches like OSS, with its radically different concepts of both ownership and development process, demand gradual adoption with a clear vision of where the business case lies.

And that business case can be stated simply. The choice is to establish an alternative ‘automotive-born’ software environment with the potential to succeed where the standard model is clearly failing – or to run out of time and money, soon.

Dr. Matthias Kempf


Christian Kaiser


Dr. Matthias Kempf

Dr. Matthias Kempf (1974) was one of the founding partners of Berylls Strategy Advisors in August 2011. He began his career with Mercer Management Consulting in Munich, Germany, in 2000. After earning his doctorate degree and further consulting work at Oliver Wyman (formerly Mercer Management Consulting), he joined the management of Hilti Germany in 2008. At Berylls, his area of expertise is new mobility services and traffic concepts. In addition, he is an expert in developing and implementing new digital business models, and in the digitalization of sales and after sales.

Industrial engineering and management studies at the University of Karlsruhe, Germany, doctorate degree at Ludwig Maximilian University, Munich, Germany.

Christian Kaiser

Christian Kaiser (1978) is partner and Head of IT at Berylls Group with focus on software and connectivity. He began his career with Daimler AG in 1997. He brings 26 years of industry and consulting experience in the automotive industry serving in various executive roles at OEMs and software companies.

At Berylls, his area of expertise is software development & architecture, digital business models, digital operating models and ADAS/AD.