Success-Plan Services
Thomas Bloor
In my discussions with traditional OEM automotive executives, three topics consistently emerge as primary concerns: software-defined vehicles, AI, and competition from agile manufacturers in China. From these conversations, one key underlying question is top-of-mind: How can my organization be more effective with the new technology? Or more specifically, how do we make software work?
Let’s take a deeper dive into these current topics, starting with software-defined vehicles (SDVs).
Software-Defined Vehicles
The term SDVs gained traction in 2023, and then in 2024, the industry largely backed away from jumping straight into central compute architectures within a single generation. Now we are seeing the major traditional OEMs adopt an incremental approach to migrating architectures across several generations of automotive architecture.
The reality is that 'software-defined' is not necessarily tied to the underlying physical architecture. However, this challenge still involves pivoting from a hardware focus to a software focus, as well as finding ways to rapidly integrate across heterogeneous systems.
This approach was very much in evidence at last month’s CES 2025 event, during which silicon vendors demonstrated ADAS and Infotainment solutions in line with consolidated architectures. Silicon vendors are well placed to speed the migration to an ultimate central compute architecture, using a strong, flexible and scalable data communication backbone. Data communication is the key to interoperable data exchange and software performance in SDVs. To accomplish this, many are turning to the Data Distribution Service (DDS™) standard, which enables increased software reuse across generations.
Artificial Intelligence (AI)
Another key focus for manufacturers is AI, which will continue to define automotive conversations through 2025 and beyond. Jensen Huang’s keynote at CES got everyone talking about how “Physical AI” will perceive, understand and perform complex actions in the real world. OEMs are understandably eager to improve the in-car user experience by adding Generative AI assistants. Meanwhile under the hood, AI will increasingly control algorithms for autonomous vehicle functionality.
For software teams, Generative AI and Large Language Model tools will help to speed the development process. They are using tools such as the RTI Connext Chatbot and CodeGen tools to accelerate DDS-based development, because these AI tools enable engineers to not only get instant technical support, but also rapidly generate code for DDS data models.
Across the broader automotive landscape, chiplet manufacturing is also an intriguing vector. As chiplets are small processing cores that can be assembled into a multi-chip package, custom or semi-custom chips can be created and tailored to a specific application, such as autonomous driving.
Mercedes Benz, for one, is investing heavily in chiplet architectures at present. They have recently announced a slew of patents in this area, including ones for chiplet to chiplet communications and failover for safety critical systems. Their stated desire? To drive this technology to become an industry standard. Such a move could enable AI and autonomous drive systems to be realized in the vehicle at a lower cost and at a higher level of reliability than conventional chip architectures.
Traditional OEM executives are also highly focused on the growing challenge from electric vehicle (EV) manufacturers in China. At CES, Zeeker displayed its sleek, low-cost production models, all of which are software-defined from day one. Other manufacturers from China showcased some truly amazing technology – everything from 100kWh batteries that can charge in 13 minutes to impressive new motors and performance models capable of going from 0 to 100km/h in just over 2.0 seconds.
New Competition
While many people marvel at the rapid innovation achieved by China’s manufacturers, it’s important to remember that the majority are startups that started with a blank sheet of paper, enabling a software-defined architecture to be built from the ground up. In other words, these architectures did not require integrating legacy technology, nor leveraging past investments. For this reason, many leading OEMs in China such as XPENG are all-in on leveraging DDS to accommodate rapid software exchange across the vehicle. They are thus able to create production vehicles faster with DDS, which allows them to build a platform that is not only abstracted from the underlying hardware with a suite of applications, but also one that can be conveniently reused across their product portfolio. It’s a great demonstration of making software work.
What about traditional OEMs who aren’t starting their vehicle architectures from scratch? Fortunately, the same principle applies. Using DDS can help to accelerate the transition to a software-defined system by abstracting the application layer from the underlying hardware. This enables reuse of application code across vehicle lines and platform generations, so that developers can focus on application development and avoid the time-consuming task of enabling their code in a hardware-defined platform. Therefore, a software-defined, DDS-based platform can enable faster migration across generations as the hardware consolidates and shifts in the future. Call it future-proofing, or call it embracing the principle of data centricity – either way, the benefits are clear.
In fact, the advantage of building on a data-centric, DDS-based architecture multiplies when you get to the promised land of software-defined vehicle production. This is where scalability, reliability, safety and security come into play, in addition to the advantage of being able to simplify system interoperability. The result? Future features and applications downloaded to the car ‘announce’ themselves on the DDS databus, while the DDS discovery process ensures that they receive the data they need to perform.
In a world looking to ‘shift left’ and speed time to market, more and more development work is moving into the cloud. Here too, the DDS-based framework can simplify and accelerate communication. DDS enables developers to develop, build and test applications in a digital twin environment, and leverage and re-use their simulation code for production programs. This improves workflow optimization by eliminating redundant coding efforts, as the DDS code can be re-used at every step of the development process.
As I continue my conversations with automotive leaders, I anticipate much discussion around the opportunities and increased competition resulting from this industry shift toward intelligent, software-defined systems. The architecture migration of traditional OEMs to central compute and a software-defined end goal will continue. By deploying a DDS-based approach, OEMs can accelerate time to market by migrating software across architecture generations to embrace the software-defined future.
To find out how RTI software helps support today’s SDV manufacturers, please read about RTI Connext Drive here.
About the author:
Thomas Bloor is the Director for the Automotive Market at Real Time Innovations (RTI).
Prior to RTI, Mr Bloor held several positions in the automotive industry including VP of Sales for Acerta, an AI/machine learning company, and Senior Director, Americas Automotive Business at BlackBerry QNX, where he was responsible for building, developing and maintaining relationships with automotive OEMs and Tier 1's.