Benefits and Challenges of Software-Defined Vehicles

The automotive industry’s steady transition to “software-defined vehicles” could offer new entrants advantages against established incumbents.

----- Transcript -----

Lee Simpson: Welcome to Thoughts on the Market. I'm Lee Simpson, Head of Morgan Stanley's European Technology Hardware Team. 

Shaqeal Kirunda: And I'm Shaqeal Kirunda, from Morgan Stanley's European Autos Team. 

Lee Simpson: On this special episode of the podcast, we will discuss the evolution of autos in the direction of software defined vehicles. It's Thursday, 19th October at 10 a.m. in London. 

Lee Simpson: Cars are in the process of transforming from electromechanical terminals to intelligent mobile devices, and we think the emergence of software defined vehicles or SDVs, is a sign we're approaching the car smartphone moment. The migration to SDVs is part of a broader transformation in autos that could even redefine the economics of the car itself. The implications for this are deep and far reaching. So Shaqeal, what is an SDV and how is it different from most cars on the road today? 

Shaqeal Kirunda: Thanks Lee, so most people are aware of one of the global megatrends in autos to transition to electric vehicles, was less well understood as a transition to the software defined vehicle. An SDV can be defined as any vehicle that manages its operations or adds new functionality, mainly through software. What that actually means to the consumer is a car that features an operating system which is upgradable over the air, not just for apps and infotainment of a whole software upgrades, safety improvements and new functions such as autonomous driving. So for a future SDV, the functions will be defined by the software and not the hardware. This dynamic mirrors how we use apps and software in phones today. Lee, how does this change the whole architecture of the car? 

Lee Simpson: Yeah, I think computing needs to change. We've seen that in other devices before and here for the car, it's transitioning really from this distributed area of lots of independent microcontrollers or simple chips in the car,ix notes towards something a little more orchestrated or a centralized compute is perhaps the best way to think of this. Now, there will not be a set path. Different OEMs and different platforms will be built along different lines, a logical path, a physical rewiring path. Some will move through domain clusters, others will move to zonal compute. But in the end, the journey will be the same. We'll move to this sort of server on wheels type of architecture, at least from the point of view of compute. And along the way will introduce new players to the automotive space, those larger chip makers who are champions in the systems on CHIP or SOC environment today. And perhaps for them they'll be attracted to this perhaps large silicon TAM that we'll see in the car. We think perhaps $15 billion of extra semiconductor building materials by the end of the decade. So with that in mind, in essence, we think the evolution towards SDVs involves a decoupling of the hardware and software in a vehicle. So, Shaqeal, where are we in this complicated process right now? And what are some of the paths to the future? 

Shaqeal Kirunda: Interesting question. We're certainly seeing different rates of progress. The key distinction here is between legacy players and new market entrants. New market entrants have embraced the transition to both EVs and SDVs. Through this they can offer over the air upgrades and safety features as well as new functions, creating new software based revenue streams. Legacy manufacturers have taken note of the major transition they're facing, but as incumbents have taken slightly longer to put this into action. Whereas the new market entrants started from scratch, the incumbents are redesigning manufacturing processes they've been executing on for years. They are making progress however, the first newly designed software defined vehicles are scheduled to be released between 2024 and 2026. But if we take a step back for a moment, pandemic caused a major disruption to the semiconductor supply chains that are so central to the auto industry. How will the migrations to SDVs change the use of and reliance on auto related semiconductors? 

Lee Simpson: Well, I think from a reliance perspective, we've already seen that in cars. There's quite a considerable reliance on those microcontrollers we've mentioned already. But if anything, this will increase. And I think you'll see that a lot of the main consideration of how a car works running through this myriad of new semiconductor chips. I think the key consideration here, however, is this is a safety critical environment and this is not something that compute is normally structured for. If you take, for instance, the cloud or even your mobile phone, the consideration here is far different. Sometimes it's about performance as in the cloud. Sometimes it's about low power or power efficiency as in your smartphone. Here the paramount feature is safety criticality. And so I think silicon here will need to have real time compute. So zero latency in its and its ability to deliver a decision maker to the decision to the driver and will also have to be secure. So I have to ensure that no new threat surface is introduced to the safety critical vehicle. So with that all in mind, what are some of the benefits of SDVs for both the auto industry and the consumer? 

Shaqeal Kirunda: Thanks Lee, the benefits for the auto industry are clear. Legacy OEMs face competitive threats from new entrants focused on SDVs. If legacy players don't transition towards SDVs on time, they will continue to lose global and local market share. Of course, the opportunity for OEMs is that the new software features could come with new software margins. Potential benefits for customers centered more towards new features and residual value. New features could be anything from safety improvements based on driver data to completely new apps from third party developers, downloaded straight to the car. Also with much better software comes much better data collection. This opens the door to predictive maintenance and improved reliability, which reduces repair costs and supports residual values. The question with all these benefits is whether customers will really value them. It will take a change in consumer behavior to shift from buying a car with all functions upfront to buying new functions later down the road. So clearly there are also a number of challenges on the road to adoption. Lee, what are some of the hurdles and downside risks of right now and looking towards the future? 

Lee Simpson: Well, I think the key thing here is software testing. This is something that, again, really leans on that safety, criticality environment of the vehicle. So before you can introduce software into a car, probably needs to be certified as safe for this environment. Now, that's a non-trivial task to overcome. Creating a certification process needs a Cross-Industry  agreement and needs someone to drive this through, and probably someone also to drive some standards that will impact in the hardware space equally as well. This will all have to be done with commercial considerations as well, so you'll have to ensure that this is consistently delivered so that the user experiences is the same car after car. This will ensure that the OEMs can deliver on their specs and the SDVs themself will start to grow as a possible value proposition for them. So finally, Shaqeal, what are some of the key milestones that investors should watch for in the migration to SDVs? 

Shaqeal Kirunda: Absolutely. Over the next few years, we'll start to see legacy players release their own version of newly updated, fully software defined vehicles. We're still at the early stages and it may take some time, but I expect we'll see further partnerships with start up automotive software players as legacy manufacturers recognize they are the best app developers. OEMs may also open their app stores to third party developers and invite them to create new applications for consumers. We've seen this with everything from smartphones to blockchain, and this could also be important for SDVs. Now, once things really take off, OEMs are sharing data and software based revenues. The key focus here will be the split between embedded and standalone revenues, i.e. those software features sold at the point of sale versus those sold during the life of the car. 

Lee Simpson: Thank you, Shaqeal. Thanks for taking the time to talk to me today. 

Shaqeal Kirunda: Great speaking with you Lee. 

Lee Simpson: And thanks for listening, everyone. If you enjoy Thoughts on the Market, please leave us a review on Apple Podcasts and share the podcast with a friend or colleague today.