The software–defined car is growing rapidly and will impact many parts of the automotive industry, including technology, market segments, business models, regulations, legislations, and all the players from OEMs to suppliers.
A recent column looked at software–defined vehicle complexities and addressed many factors that impact automotive software in the software–defined vehicle era. It also provides perspectives on current and future trends.
The figure below shows an overview of important factors that are influencing the software–defined vehicle market. The impact is segmented into four categories: technology advances are in black at the bottom and legislation and regulation are on top in green.
Software–defined vehicle forces have two categories — blue boxes on the left for the development of the necessary software platforms and red boxes on the right for the customer use phase of the vehicle.
Semiconductor technology advances remain the most crucial factor in expanding software use. Many of the semiconductor improvements are now coming from on–chip and multi–chip connections and packaging advances ranging from System in Package and 3D chips to chiplets. Chiplets are receiving increasing attention with the announcement of an open standard called Universal Chiplet Interconnect Express (UCIe). UCIe was developed by Intel and donated to a new consortium that includes many of Intel’s competitors.
AI technology is also having a significant impact on automotive software. AI training is key to developing applications for user interfaces such as speech to AV software platforms. AI inferencing is a necessary component for these AI applications to work in vehicles. Moving forward, AI chip acceleration will be necessary to lower AI system cost and power consumption.
All vehicles will be connected with software–centric systems now being the most important driving factor for the necessity of this connect. An earlier column provides additional perspectives on this topic.
Battery electric vehicles (BEVs) are in take–off mode and progress is moving much faster than expected courtesy of Tesla’s growth and impact. The battery management system is a key software platform and will increase in importance as BEVs become a backup to replace house electricity and power sharing with the electric grid.
Legislation & regulation
Legislation and regulation tend to lag technology development and this pattern holds true for automotive software as well. Additionally, many legal and regulation activities that are not focused on software indirectly impact automotive software.
Vehicle data use is getting increasing attention with some regulations already on the books. There are local, regional, and country–specific rules that are constantly evolving. A recent column on Otonomo’s connected car activities shares good perspectives on the difficulties of keeping up with all the vehicle data regulations.
Road safety is defined by legacy laws that are undergoing significant changes from ADAS and NCAP–driven amendments. Big changes will happen when the road use legislation for AVs take effect in the next few years.
AV regulations are also coming from other organizations including ISO, SAE, and IEEE, which specify technical and safety requirements. Examples are ISO 21448 (AV functional safety), UL 4600 (safety for driverless AVs), and IEEE P2851 (AV crash avoidance).
Oversight is also required to give AV driving permits and track safety performance and expected lifetime advances in driving skills.
Software–defined vehicle development phase
The development of all software platforms required for software–defined vehicles are undergoing major changes. This includes new and emerging technologies such as electronic system architecture, software architecture, development tools, testing systems, and more.
Electronic system architecture is focused on domain ECUs and Ethernet–based networks. These changes will provide a hardware base for safety–centric electronic systems that the software architecture can build on and represent the growing trend of moving away from the CAN bus.
The core of the software architecture for this particular use case is now a layer of interconnected software platforms that use a service–oriented architecture (SOA). Small SOA components called microservices are used to lower the complexity of the software building block.
Automotive software development was traditionally done via SDKs (Software Development Kits) from a variety of suppliers. In the last 5+ years, open–source SDKs became the leading software development tools with Integrated Development Environment platforms such as Eclipse as the leaders.
Cloud platforms for software development have increased in automotive software development with AWS at the forefront. Large infrastructures and ecosystems are expanding and becoming increasingly important for auto software platform creation.
The next technology to impact software development are AI platforms that will generate software code. This technology is newly emerging and limited and will need more testing and advances before it will be used in automotive software; especially software with functional safety requirements.
Cybersecurity and OTA have been identified as key technologies and have received legislative attention. Both embedded and cloud–based cybersecurity software are integral parts of creating automotive software.
Embedded cybersecurity software is needed to protect most ECUs, particularly ECUs using wired or wireless connections. Domain ECUs have access to many other ECUs and require their own cybersecurity hardware and software.
Cloud cybersecurity platforms are key to protecting fleets of vehicles as they provide real–time attack detection for everything from logistics fleets to OEM fleets. Cloud platforms provide situational awareness for a fleet’s cyber health and threats. That platform is often called the Security Operations Center.
Over–the–air (OTA) software update capabilities have become a core function that auto OEMs require to manage their software–defined vehicles. OTA provides the core technology that makes the car operate over its lifetime and will create multiple new revenue streams.
However, OTA platforms are becoming more than just OTA. Additional connected car functionality based on OTA technology is rising sharply and nearly every OTA platform can be expanded to add other applications with data flowing to and from cars.
Software–defined vehicle use phase
The 15–year use–phase of most cars will see significant improvements from software–defined functionality for the customers, OEMs, and suppliers. For customers this is due to the potential of upgraded functionality during the lifetime of their vehicles.
For OEMs this is due to the potential of new business models that are based on software platforms combined with OTA updates and connected car functions. Software has historically been a cost center for auto OEMs due to expensive development and lifetime debugging and maintenance. An OEM software profit center is emerging based on various business models from Software–as–a–Service (SaaS) and cloud platforms.
OTA and cybersecurity updates are becoming SaaS products that every vehicle will use throughout their lifetime. The payee will vary between OEM and customer depending on what is updated.
Data–as–a–Service started with free passenger content via radios, which was truly the first infotainment category. Streaming content is now popular via smartphones and web content sites. When future AVs are common, the content segment will get a big boost as video content can be consumed by all passengers.
Data from ECUs is growing in value to many companies. Much of this data monetization has been developed by independent companies that receive data from OEMs via revenue sharing deals. Otonomo and Wejo are examples of data monetization startups. While data to ECUs has potential, most of it is counted as OTA updates.
Function–as–a–Service includes a few categories with paid functional OTA updates as the key growth factor. Tesla has shown this potential and other OEMs are working to expand their portfolio.
Remote diagnostics and prognostics, which are part of the telematics systems, can also be part of this category. There is likely to be an increasing number of business transactions done via vehicle systems and software platforms, including purchasing transactions that could also be done via smartphones.
Mobility–as–a–Service (MaaS) is in the early stage of use but is expected to create lots of future value through its four listed segments. Autonomous trucks and robotaxis in particular show great promise in the MaaS space.
The auto industry is on a path to reap the benefits of software–defined vehicles that will last customers, OEMs, and suppliers for the lifetime of the vehicles.
But the development of the software–defined vehicle is complex, time–consuming, and expensive and requires many resources that the auto industry does not yet have. To add complexity to these tasks, the development phase must also include the software platforms that manage and update all the software functions during the vehicles’ 15–year lifetime cycle.
The use of cloud platforms to develop and manage this software is becoming crucial for all OEMs and their suppliers. When done right, the auto OEMs have the potential to become increasingly profitable enterprises.
For vehicle purchasers, this may sound like bad news due to higher expenditures. However, the software–defined vehicle of the future will provide more lifetime benefits and value. If higher customer value does not happen, the OEMs will remain at current profitability levels or be replaced by better software–defined vehicle suppliers.