Mobility Innovation Summit 2025
"The future of the vehicle lies in the code!"
Interview with Georges Massing, Vice President MB.OS Automated Driving & E/E Integration at Mercedes-Benz AG, about the concept of the Software-Defined Vehicle (SDV) and how software can redefine the automobile
Interview with Georges Massing, Vice President MB.OS Automated Driving & E/E Integration at Mercedes-Benz AG, about the concept of the Software-Defined Vehicle (SDV) and how software can redefine the automobile
Mr. Massing, in your opinion, what are the drivers in the development towards the Software-Defined Vehicle (SDV)?
Massing: Modern vehicles are highly networked systems with a multitude of control units and sensors. Instead of further increasing this complexity with additional hardware, software enables a more efficient, flexible, and future-proof architecture. It reduces dependence on specialized hardware, enables continuous development, and makes the vehicle more adaptable than ever. These are precisely the advantages driving the development of software-defined vehicles.
Furthermore, our customers are accustomed to an intuitive, seamlessly connected digital world, and they expect the same from their vehicle. A modern car should integrate into their digital ecosystem as seamlessly as their smartphone or smartwatch. This means a car that "thinks for itself", anticipates, and feels as if it were made specifically for them.
Production also benefits significantly from the SDV concept: The decoupling of hardware and software enables more efficient production, as the vehicles can be developed and produced independently of the final software. Just-in-time software delivery ensures that the latest software version is installed shortly before or even after delivery: flexible, needs-based, and always up-to-date.
Conclusion: The future of the vehicle lies in the code!
To what extent do vehicle owners benefit from technological developments?
Massing: A software-defined vehicle brings numerous advantages for users, three of which I would like to briefly introduce:
First, over-the-air updates: An SDV doesn't age, it matures. Like a fine Italian wine that improves with time, the vehicle also continues to evolve after purchase. New features, performance improvements, and security updates can be installed at any time - without a visit to the workshop.
Second, the personalized user experience: Thanks to software-based architecture, the vehicle can optimally respond to the individual needs of the driver - from intelligent assistance systems to personalized infotainment functions.
And third, safety, efficiency, and sustainability: Intelligent software optimizes energy management, increases the range of electric vehicles, and ensures predictive maintenance. This ensures that vehicles remain not only safe, but also efficient and sustainable.
Where do we as the German and European automotive industry stand in terms of SDV in a global comparison and what are the most important success factors?
Massing: When it comes to software-driven architectures, Chinese companies dominate, followed by US companies. The key to success for Germany and Europe lies in the following measures:
First, building our own semiconductor and cloud infrastructure: Greater European independence in chip and cloud development is necessary to remain competitive.
Second, the development of and with artificial intelligence: The integration of artificial intelligence into vehicles and our work processes must be further advanced. AI enables customers to experience more flexible and efficient control of functions for an even more personalized user experience. And the company benefits from shortened development phases thanks to optimized development and validation.
It also involves cooperation between industry and politics: Regulations should be designed to promote innovation in order to bring technological advances into series production more quickly. The European Artificial Intelligence Regulation already lays promising foundations for data-driven development with AI. Additionally, further concrete measures and decisions are needed in areas crucial for the future, for example, the establishment of our own chip production facilities and high-performance cloud infrastructures.
How do you and your team work together to be successful in developing software-defined vehicles?
Massing: Our approach at Mercedes-Benz is based on three key principles:
First, our own development platform: We have created a fully integrated development environment that enables fast, AI- and data-driven, agile software development, testing, and deployment.
Second: full control over the software stack (editorial note: see explanation below the text). This means that our in-house developed Mercedes-Benz Operating System is designed from the ground up as Mercedes' own software architecture. This ensures optimal control of the vehicle software by our engineers according to customer needs.
And third, hardware agnostics (editorial note: see explanation below the text): A modular architecture allows software solutions to be reused across different vehicle models and generations, regardless of the hardware. These advantages make us efficient and accelerate development cycles. Our goal is to combine an automotive "Silicon Valley" mentality with Mercedes-Benz engineering expertise to remain a technological leader.
In a session, you discussed how software is redefining the car itself with Christian Salzmann (BMW Group), Detlef Zerfowsky (ETAS), and Dr. Elmar Pritsch (Deloitte). What common challenges do you see?
Massing: Innovation leadership is no longer determined by horsepower and displacement, but rather by intelligent algorithms and powerful computing in the age of new electric powertrain technologies for maximum range despite maximum performance and software-defined, computationally intensive vehicle functions. In a competitive environment, speed and innovative strength in development are crucial to avoid offering outdated technology tomorrow.
A fundamental shift is emerging: Instead of adapting software to the hardware, software now defines the hardware - a paradigm shift in the automotive industry. Neural networks, high-resolution gaming and streaming in vehicles, and the bundling of complex software functions require water-cooled supercomputers and powerful graphic processing units optimized for software algorithms. One example is the development of neuromorphic chips (Editor's note: see explanation below the text) for computing of neural networks in vehicles.
Three key challenges emerge:
Traditionally, vehicles are the result of the interplay of numerous suppliers, each with their own, often proprietary, systems. However, the transition to a software-defined vehicle requires a consistent, integrated architecture - from sensors to the cloud. It's all about integration rather than isolated solutions. While software companies release updates weekly, the automotive industry demands the highest safety and reliability standards. The challenge lies in reconciling agile, innovative software development with strict regulatory requirements.
Another challenge is the race for digital sovereignty. As tech giants increasingly penetrate the mobility sector, the automotive industry must assert its own innovative strength. The question is: Will the car of the future be developed by traditional car manufacturers - or largely by software companies?
One thing is certain: whoever masters the software defines the driving experience. The automotive industry must therefore not only expand its technical expertise but also redefine its role in a changing ecosystem. Mercedes-Benz has decided to become a software company and become a leader in the development of vehicle software.
Editorial notes:
A "software stack" (also called a solution stack or simply a stack) is a stack of software layers consisting of various software components that build on one another to act as a platform to support the execution of a common application.
The term "hardware-agnostic" means that a software or system functions independently of the underlying hardware platform. This is achieved through abstraction layers, hardware virtualization, and standardized interfaces in the software, among other things.
A "neuromorphic chip" is a microchip that is built based on the model of natural neural networks.
Photo credits: VDA/Ch.Lietzmann/M.Schwarz




