Changing Roles of Automotive Software

Cars are rapidly becoming software-defined, and that is starting to change the automotive industry’s technology, business models, and even its very structure. Similar changes have taken place in other industries. It is now time for the auto industry to face changes and disruptions.

My definition of “software-defined” means that a large majority of car functionality is now implemented by software applications that run on the required processors, memory and sensors. Additionally, most of the functionality is defined by how well the human-machine interface is implemented in software. We are also expecting expanded software capabilities for more advanced ADAS and autonomous vehicle functionality. All of the software will require advancing connected car systems for remote updates, cybersecurity, data collection and infotainment for the occupants.

Two distinct software categories are used — embedded software clients in cars plus software-as-a-service (SaaS) platforms that are delivered via cloud software platforms. The software client platforms are dominant now in the auto industry, but SaaS platforms are on the rise and will play a much more important role in the future.

The next table is a summary of my perspectives on how automotive software will change in the next decade. In the table five category of players are listed—software developers, cloud computing providers, Tier 1 suppliers, auto OEMs and auto dealers. To show recent evolution and future projections, the columns have summaries for three time periods — 2010s, now and 2030s. The content in the table are explained below for each player category.

Click the table above to enlarge. (Source: Egil Juliussen)

Software developers
The software developers include the many companies that provide system software, middleware, and automotive application programs to the auto OEMs — usually via Tier 1 suppliers. These companies have developed software platforms for the auto industry. Each software platform is initially a “cost-center” due to high initial investment and long development time. The most common business model for software platform clients are royalty per unit shipped. The unit royalty varies by software category. The Tier 1 or OEM will also pay for tailoring the software platform to a vehicle’s hardware and system configuration, which is usually a profitable business to software developers.

After the software platform sales revenue exceeds development cost the software platform becomes a “profit-center” and usually with high profitability — especially when the platform is used in many vehicles with similar hardware systems.

Software maintenance or fixing bugs during the cars’ long life is usually a profitable business because the OEM or Tier 1 customers pay for software maintenance for multiple years. The OEM and Tier 1 customers will also pay for updating many software platforms — usually when a car model is updated.

The business dynamics described above have been used for over a decade but is now changing quickly. The rapidly growing connected car has opened the door to remote over-the-air (OTA) software updates, which is creating new opportunities.

The connected car is also creating SaaS opportunities for several software segments including cybersecurity and smartphone apps that migrate to infotainment systems via CarPlay and Android Auto.

Another new opportunity is the increasing role of cloud computing providers such as Amazon AWS and Microsoft Azure. The cloud computer providers are participating in the software development phase with advanced software development tools and are especially strong in AI development. There will be more on this in the next section.

By 2030, the software developers will see large changes — both in technologies and business models. Platforms for software clients will remain important. SaaS platforms for delivery via cloud platforms will be more important and may be the largest revenue segment. OTA with built-in cybersecurity will be required for all software clients.

The software developers will need to work with all or most of the leading cloud computing providers as they will be involved with the Tier 1 and OEM’s software development effort and software product selection procedure. The software developers will also work with cloud computing providers to get a portion of SaaS opportunities.

Cloud computing providers
It is important to understand how and why cloud computing providers are becoming so important. Amazon AWS was the pioneer and is the leading player. If you want to understand Amazon, there is an excellent book called “Behemoth—Amazon Rising.” EE Times did an interview with the author, Robin Gaster that you can listen to or read: That Sounds Reasonable | The Company Eating the World | EE Times.

AWS initially developed cloud computing technology for Amazon’s internal use around 2003. AWS used a service-oriented architecture (SOA) where every software module was defined by APIs (application programming interfaces) and implemented as cloud services via the internet. This flexible system architecture worked well for Amazon’s growth and was very adaptable for other companies computing requirements. It was also less expensive and more reliable than traditional IT computing infrastructure — as much as an order of magnitude less costly and sometimes even better. In 2006 Amazon launched AWS for external use and became a major success story very quickly.

During the 2010s, AWS could offer a variety of services to multiple industries with high-tech startups and/or companies providing digital products via web deliveries — such as Netflix. The table above lists timing for the auto industry, which is later than high-tech companies.

The three services categories AWS offers are IaaS, PaaS and SaaS. Infrastructure as a service (IaaS) delivers a full computer system — including servers, storage, networking and operating software as a virtualized service. Platform-as-a-Service (PaaS) includes software development and system software resources, which makes it easier for companies to develop their own applications.

With Amazon’s investment and expertise in AI technology, PaaS-based development of AI applications on AWS is becoming popular. The ecosystem of software apps, platforms and SaaS components that is available from hundreds of software developers is another major reason for AWS growth in the last decade.

The auto industry started taking advantage of these AWS capabilities about five years ago and is now growing rapidly with much more to come. AWS has competitors with Microsoft Azure being a strong second. Azure is quite strong in the automotive industries with many auto OEM clients. There are many smaller and fast-growing cloud providers that will serve the auto industry such as Google, Alibaba and Tencent.

The conclusion I draw is that the cloud computing providers will have a growing impact on automotive software; both in the development of software client platforms and in the software use phase as SaaS and pay-per-use business models and service delivery via web infrastructures increase.

Tier 1 suppliers
The Tier 1 suppliers are very important to the auto OEMs and are the manufacturers of most of the electronics systems used in cars. Electronics hardware manufacturing is their most important business. Tier 1s are also important as software suppliers and software integrators of many software clients provided by automotive software suppliers that are included in the hardware systems.

The Tier 1s have seen the importance of software for a decade and have had some success in establishing their own software platforms through acquisition or own development. They will follow this track and are adding SaaS platforms as important parts of their future product and service portfolios.

There are multiple challenges emerging for Tier 1s both in hardware and software. The leading contract manufacturers are trying to get some of the auto hardware manufacturing business.

The auto OEMs want to have more control of their software business and are developing their own software platforms, which will have some impact on Tier 1s in the next decade.

The cloud computing providers’ business will also increasingly overlap with the Tier 1s’ software activities—especially in SaaS segments. But the Tier 1s must cooperate and use the valuable software and web ecosystems that are emerging for the auto industry.

Auto OEMs
A decade or two ago, software was a necessary “evil” to OEMs because it was all “cost-centers.” Software was hard to plan and manage and was often behind schedule. But software was absolutely required and became increasingly important and grew to be more important than hardware.

Auto OEMs have had their own software platforms for over 40 years—the application software that control electro-mechanical systems such as combustion engines, car body and  chassis equipment. This OEM core competency software have increased in code size over the last two decades but are still far from the complexity of software for infotainment, ADAS and autonomous vehicle functionality. The addition of OTA software updates and cybersecurity create another level of complexity.

Even with these challenges, the auto OEMs want and need more control of the software used in their vehicles. They definitely want to move from a pure cost-center software model to a combination of cost and profit-center software business models. Hence, they are developing their own software platforms, eyeing SaaS platforms and other pay-per-use business models.

The auto OEMs are already starting to work with cloud computing providers for a variety of application segments—including software development platforms and SaaS. I believe this trend will continue and accelerate in the next decade.

Auto Dealers
I included auto dealers in the above table, but you can see that they have little impact on the automotive software business. They currently have a profitable business in software updates and software recalls. As OTA takes hold the auto dealers will loose much of its profitable software activities. I have a question on how auto dealers will tap into future software business models such as OTA, SaaS and similar pay-per-use revenue opportunities.