Aptiv PLC (APTV) Earnings Call Transcript & Summary

Okay. Yes. If I could ask people to please take their seats because we're going to get going with the next webcast presentation, please? Great. Thank you. Once again, I'm Ryan Brinkman, the U.S. automotive equity research analyst here at JPMorgan. Very excited to have again this year, Aptiv, including Senior Vice President and the Chief Technology Officer, Glen De Vos; as well as Vice President of Investor Relations, Elena Rosman. Thanks so much for coming.

I think one of the most important developments for Aptiv since CES last year has got to be the formation of the fully autonomous driving joint venture with Hyundai. Clearly, this brings myriad financial benefits, as were visited in the last call, to margin, cash flow, returns, et cetera, and maybe you can revisit those financial benefits for the investors here. But I'd be interested, too, if you could comment on the potential commercial benefits that you might see on the third quarter call. For example, you discussed the joint venture potentially accelerating the development of advanced safety solutions for existing OEM customers. Maybe you could touch on that as well. Thanks.

Sure. I'll go ahead and kick things off on the financial side. So just as a reminder, it's a 50-50 joint venture that Aptiv announced with Hyundai, our OEM partner for autonomous mobility. And beginning in terms of just what's being contributed, so Aptiv will contribute its autonomous mobility, its Level 4, Level 5 policy and path planning software, to the joint venture. And Hyundai is contributing $1.6 billion of cash upfront at closing, in addition to $400,000 of engineering services effectively, so combined, about a $4 billion total valuation for the joint -- the JV formation. I think what Ryan is alluding to is in terms of what Aptiv has been investing in its autonomous, L4/L5, or primarily the commercial or robotaxi market, the operating costs that we've been spending about -- annually are going to be just under $200 million this year. So going forward, as part of the joint venture, the joint venture effectively is -- or the current operating expenses are recorded in our AS & UX segment. So that's $200 million of operating expenses that then get deconsolidated, moved to below the line. And effectively, going forward, Aptiv will only recognize a 50% of the profit or loss in this case of the joint venture on a -- net of tax and equity income line item. Now there will be some incremental intangibles, amortization, as it relates to the purchase accounting as in any other type of transaction of this nature, so that would be incremental. So net-net, it's probably total EPS-neutral. But what is a positive benefit to both the AS & UX and Aptiv segment margins, our operating margin, and most importantly, our cash flow. So that $200 million of annual cash spend has been funded 100% by Aptiv to date. Going forward, as part of the joint venture formation, the $1.6 billion that Hyundai is contributing to the joint venture will effectively capitalize the joint venture for the next 4 to 5 years. So it's about $200 million accretive to cash flow for Aptiv on an annualized basis, and we do expect the transaction to close sometime in the second quarter of 2020. So at this point, when we give guidance on our fourth quarter earnings call on January 30, we'll give some further clarity as to how to think about the guidance for the joint venture in 2020.

Great. Relative to the second part of the question and how does that benefit or potentially accelerate some of our ADAS development. While the JV will be very focused on the robotaxi market and its time lines and the technology requirements, we've always thought of Level 4 really being on that road map of active safety, so Level 1, 2 and then extending to 4 and 5. And so what that means is that the technology building blocks that we've developed for our advanced ADAS systems really have an opportunity to commercialize into the JV, supporting both current as well as next-generation sensors, compute, et cetera, architecture. So that's a benefit. But then also, as the JV develops its capability, its products and systems, those have applicability. As Level 2 and 2+ and 3 systems get more sophisticated, those technologies have applicability for the broader market. And that's really -- that's one of the benefits of the relationship between Aptiv and that JV. The JV has an opportunity to commercialize into the broader ADAS market, so getting really broader opportunities to monetize their tech versus just through robotaxi or fleet services.

Great. And then maybe just as a follow-up, could you touch on what has been the reaction from OEM customers since the announcement given that they will now be expected to purchase from a joint venture that's partly owned by one of their competitors? I know Volvo, Veoneer have a very similar structure with Zenuity, but just thought to ask there in case there was anything to think about.

Yes. So the discussions -- we've basically talked to all of our OEM customers and the response has been very positive and in part because the automated mobility group was always separate and very distinct from our broader ADAS organization. So it was a -- for the OEMs, it was essentially a nonimpact in any way to their business and ongoing relationships. That said, what they look at is an opportunity to -- they can either buy directly from the JV or they can buy tech from the JV through an Aptiv. And so there's a lot of flexibility in how that potential technology sale could be managed, and the OEMs appreciated that to have the flexibility. The final point I would make is, as we've always said, we needed an OEM partner as part of our Level 4 activities. This move made sense to everybody we talked. I mean there was nobody that was wondering, "Why would you do that?" And so when things make sense, they generally get received in a more positive way. And as we could explain to them, look, as an OEM customer, you can have access to what that JV does through us. It gives them some interesting options as well.

If I could, I just want to add, one of the things that also made Hyundai a fantastic OEM partner is that they share the vision for the time line and saw the opportunity in a more intermediate period for the commercial or robotaxi market really starting to evolve in that 2022 through 2025 and then scaling there beyond versus traditional OEMs where you're thinking about a Level 4 type of technology, really, in 2030, in that time frame.

Great. And I'm sure one thing everyone in Vegas this week wants to know is when are fully autonomous vehicles really coming. And I know you have fully autonomous vehicles in Vegas right now, right, both with Lyft and for your testing purposes? I guess what I'm more asking about relates to like commercial launch at scale without a backup driver. What year do you have in mind when we might see a service like that debut? And do you have any thoughts as to who might be first to launch such a service or in what geography?

Yes. I think in terms of the time line, our time line really hasn't changed. And as Elena just mentioned, with Hyundai, it gives us -- it really solidifies that vehicle partner, that vehicle supply which was a critical issue for us. So when we talk about our time line, this year, 2020 is when we would expect to start pulling drivers out of the cars, which, although a limited deployment in a noncommercial setting, is an important milestone. 2022 is when we're targeting really the first commercial deployments. Now these will be fairly limited, geofenced areas, limited service, but it is important to get cars on the road. And one of the benefits we've had with the engagement with Lyft these past few years here in Vegas is it's helped us understand deeply what you need to do, what the car needs to be able to do to be commercially viable as part of a network. And in this particular case, a hybrid network of manually and automated vehicles. And so that gives us clear line of sight to what those requirements are, how do we deploy and when and what scale. Scaling, we don't expect to really accelerate until probably around 2025. And why that's important is because there's 2 dimensions that you have to think about with an automated robotaxi, and that's the cost of that vehicle as well as its utilization. And to drive that cost down, you really need to get to a fully integrated vehicle. So the systems are fully integrated. It's not a bolt-on system or partially integrated. In 2025, working with Hyundai, that's when we expect to get to that fully integrated platform that has a low-enough cost so that it can support scaling. It has the OEM partner, and then we'll have the ODDs mature enough to actually go to many, many markets. It will -- the practical reality is, so we'll pick markets that are favorable for robotaxis. And so you're going to see that in markets like -- not to say that these are the selected markets, but markets like Vegas, markets like L.A., Arizona, Florida, Singapore, things like this where you have environments and infrastructure and municipalities and all of those things supporting the deployment. In terms of who gets there first, I mean, that's a great question. I think there's been a lot of dynamics around timing for a number of the players. I wouldn't -- if I could predict that, I'd be doing something else. But I think we're very comfortable with our time line.

Great. And then just relative to the technologies that are going to get us there to the fully autonomous driving future. Can you talk about which sensors you think will be needed? How have the requirements may be changed over time? And with the new developments in the industry, how many cameras, radars, LiDARs on a vehicle, for example?

Yes. So in terms of the types of sensors that are needed, and it's really important as we think about this, the perception system of the vehicle is what really limits that vehicle's ODD or its operational design domain, what it can do. To the degree that it can see, perceive the environment around it and predict what's going to happen, that really defines, okay, how is that vehicle? What can it potentially do? So the general philosophy you see in the industry is let's make sure we have the best perception systems possible. We believe you need radar. You need vision. You need LiDAR operating in 360 degrees around the vehicle, both for maximizing the perception capability but also for fail-operational capability and redundancies. Every perception modality has strengths and weaknesses. And so a really great question is, well, how many of these things do you need. So we're all trying to solve for what's the minimum viable set of sensors and the mix between those modalities. And the more we can do with radar, it means the less I have to do with LiDAR, vision. And so that's still in flux. And as those sensors get better, as those sensors get less expensive and easier to install in the vehicle, you're going to see that mix change. So I can't really give you exact numbers at this point in time. What I can say, at the end of the day, you're going to need to have a full coverage of that vehicle. And depending on the operational design domain, you may need to have very long-range sensors and maybe a focus on short range. And every -- so your service environment is going to dictate the sensor mix as well. But that's really kind of as we have thought about it, that's our current strategy.

Okay. And I'm curious, too, if you see a role for still other types of sensors in the car. So for example, I don't know, ultrasonic, night vision, thermal, et cetera, to complement the more often discussed trio of vision, radar and LiDAR.

Yes. Where those sensors can unlock use cases that are meaningful for the service, you're going to see people adopt those sensors, and that's kind of the way to look at it. If by using a thermal imager or a night vision, I can unlock certain use cases, I can drive in certain areas that without that I simply can't, then I'm going to need to adopt that sensor because at the end of the day, you have to be able to manage all of the different use cases within your service area. What we found to date is that when we think about the initial deployments, we're not using thermal imaging, and we really don't use ultrasonics either. Park aid, I guess you could. But -- and we're not using night vision at this point in time. Now as their costs come down, they become more interesting because they have the potential of unlocking certain use cases that could be at a point in time absolutely necessary. Today, we don't see that as the case. But as you look down the road and you think about a car that can drive anywhere, it's going to see a lot more -- there's going to be a lot more demands on that vehicle in terms of the use case complexity, and you may see the need for some of these different types of sensors coming into play.

I suppose the next question to ask, how much is all this going to cost? Are all these radars, cameras, LiDARs, also the fusion software, too, what do you think the total cost is now? Actually, I think you know what the total cost is now because you create the makeup of a fully autonomous vehicle. I don't know if you shared what that number is. But also what will that cost be when we do have commercial launch? And how could it trend downward over time? What are the catalysts for trending it downward over time in order to facilitate proliferation?

Yes. Generally, as I mentioned earlier, there's 2 key variables associated with commercial viability. One is cost of the platform of the vehicle, and then the other is the utilization of the vehicle, which is a direct function of its ODD or its capability. So cost is very important to us. And today, when you think about these systems, they're bolt-on systems. So the BMWs that we're riding around here in Vegas are extremely expensive. And as a result, not -- they're not a viable option. As we think about launch in 2022, though, what we're targeting is somewhere in that neighborhood of around $125,000 for the vehicle hardware, all the tech that gets added into that vehicle that has to, on top of the vehicle, I mean, including the vehicle cost. And you've got to be in that neighborhood. And so there's pluses and minuses associated with that, depending on what you want that vehicle to do and what you expect it to do. But then you have to hit a certain utilization rate to make the math work. And that's -- that number has been looked at from a variety of different lenses, and that's an achievable number. So I think when you think about the -- where the industry is for robotaxis, commercial fleets and initial deployments, noncompetitive markets because you're not going to have a whole bunch of these fleets, that's a viable number. Now what's going to happen, though, is the cost of that system has to come down dramatically because that will be one of the key barriers to scale. And so how fast that comes down really relates to how integrated you can make the system into the vehicle architecture. And so you're not adding on and getting redundancy through duplication. You're integrating it. You have a much more efficient cost basis. So as sensor costs, if you think about radar and vision, those are already mature technologies with low cost. So LiDAR, that's the one that has to obviously come down significantly. Compute has to be optimized. And then that leaves you with the software, which is really more in a value-based pricing. What value are you delivering with that platform? But as we've looked at those economics, one, I think they're very achievable; and two, we think they're very attractive. And so when does that happen? That's really that 2025 time frame when we see scaling really taking place.

Great. And we've been talking about radars, cameras, LiDARs. Obviously, these are all hardware components. They do include a significant amount of sensor-level software to, say, of the fusion software, too. To what extent can Aptiv be thought of as a software company? How many software engineers do you have? What percentage of your value add do you think in the future could be derived from software versus hardware? And what kind of implications might that have for financial metrics like margin or returns?

Yes. So in terms of just the raw numbers, we have about 20,000 engineers overall. 7,000 of those engineers are focused on software, the different aspects of software development. We do everything from the very base software that sits inside of a sensor or sits inside of compute to the sensor fusion for our ADAS systems, working with Mobileye, who's our partner in vision, and then also the future-level development, so whether that's back -- blind spot detection or Lane Keep Assist, adaptive cruise control, cross-traffic alert, all of those types of features. So we do all of those things. And so if you think about Aptiv, a significant portion of what we provide today is actually software. In fact, on many of these ADAS programs that we talk about, these kind of satellite architecture programs that we're developing, the vast majority of the resources applied to those programs are software engineers. Over 80%, generally speaking, are the engineers on these programs, where there's hundreds of engineers who are software. So today, though, we still are operating primarily in a unit sales-driven type of an environment where the OEMs buy the system, hardware bundled with that software together. As we think about where we're going with more advanced architectures where software is abstracted from the underlying hardware, we see a great opportunity there to capture more value in the software layers because it's going to be managed. It's going to be developed, and it's going to be basically integrated separately from that hardware. Today, they're very much entangled, very much embedded. Where we're going in the future is where that software is abstracted, which means there's models, not only for the initial licensing of the software, which we're starting to do now, but also for recurring revenue streams on the maintenance of that software, the ongoing improvement of that, the introduction of new features. And because what really happens is if you think about software in the vehicle today, not only is it tied to hardware. It tends to be tied to the model year launches of those vehicles. You're developing a new system for a given vehicle, and so everything is around that vehicle. What we're talking about in the future is where that's decoupled. So the software, just like your phone, is developed and improved and grows over time independent of those underlying hardware platforms. And when you do that, you now have the opportunity to have -- open up many new business models in that software. We think that those are all positive from a margin expansion basis. How that all plays out, I mean, obviously, we're early into that development. But for us, we see that as one of the really important areas for us being able to create value in those software layers.

Okay. And we've talked a lot already about Level 4, Level 5. Maybe if you could just give us an update on Level 2, Level 2+, you sometimes refer to it as active safety or semi-autonomous by the investors. You've noted an increase in interest recently by OEMs for these technologies. Are you able to kind of dimension that for us at all? What is the potential benefit to Aptiv? I'd be interested, too, if you could -- do you have any thoughts on what is driving that increased interest? Is it more like the consumers are demanding it, and they're willing to pay for it, that's creating excitement amongst the OEMs? Or is it, I don't know, more regulatory-driven or maybe an attempt by automakers to differentiate their products from other automakers?

Yes. Maybe I'm going to reverse the order of the questions. I'll start with the second one first, and then we'll kind of back into the -- what does it mean for Aptiv. So active safety or ADAS, advanced driver-assistance systems, is at a tremendous period of growth right now. And you'd mentioned the 3 -- kind of the 3 forcing functions of the -- factors. One is, and this is really an important one, is the consumer awareness. Consumers are becoming aware of these systems. They're starting to understand how they work. And what we're finding is that they -- there's a strong pull for these systems because they're seeing the benefits. And active safety systems are incredibly sticky, so it's a 95% rebuy rate. So if you've been driving a car that has blind spot detection and adaptive cruise control, has cross-traffic alert, has all of these features, the likelihood of you buying your next car or leasing your next car without those features is basically 0. In fact, what you're really going to want to do is, see, I want the next generation, I want more of that. So what we're seeing from the consumer-facing side is a very strong pull and a very sticky pull for these technologies. For the OEMs, it's a great option to, a, differentiate, but also from a margin standpoint. These are higher-margin options under it. So what we sell to the OEM is typically marked up many times. So for them, selling a car with active safety content is very good for their financials. And then, finally, from the regulatory and the safety agency standpoint like NCAP, New Car Assessment Programs, the regulatory bodies, there is a strong focus on expanding ADAS capability, vulnerable road users, bicyclists, pedestrians and continuing to see the deployment of those systems as well as expanding the capability of those systems because they're seeing the benefits. We're getting now data from the market from a lot of units in the field, millions of vehicles in the field, that show they are impactful in terms of reducing accidents, saving lives, reversing the trend of the fatalities per 100 million miles being driven, counteracting driver distraction, all of those things. And it's not just the classical ADAS features that we all think about. It's now things like interior cabin sensing, driver state monitoring. There's just a whole host of other adjacent technologies that we're developing that we can basically layer on top. So for us, it's just a really, really -- we see it as a very attractive market and one that we're happy to play in. I think -- now Aptiv has been very successful in this space, mainly, I would say, largely because, one, we have the building blocks, right? So we've been in radar now for 20-plus years since 1999, our first production launch with Jaguar, our adaptive cruise control. We have a great partner with Mobileye on the vision side. But we also have the combination of putting that all together, what we call satellite architecture, where we take basically the intelligence or the processing out of the sensors that make the sensors very small, very easy to package. They bring it all back into a central domain controller or ADAS controller. So all the software is in one package now. It makes it much easier for the OEMs to develop, to apply across multiple vehicle platforms, and it saves them money. And so that's been extremely successful for us. And our ability to do that systems integration to manage all that complexity for the OEMs is one of the keys. And I think Elena can speak a little bit to the results, but that for us has just been a great story.

Yes. I guess I would comment just on the results. So our active safety bookings over the last couple of years, going back maybe 3 years, were somewhere around $1.5 billion a year. We're now booking in the $4 billion-plus-type range in terms of annual bookings. So not only -- I think this -- it's sort of how we've defined the sweet spot for Level 2, Level 2+, Level 3-, Aptiv is perfectly positioned, especially in the area of smart satellite architecture, to solve a lot of those challenges for the customers, for our OEM customers and deliver on that consumer functionality that Glen referenced. McKinsey did a study where, today, the majority of active safety systems still are option packages. And their estimation on rebuy rates, it's the highest rebuy rate, obviously, in the industry at roughly 95%. So that's not changing. And what we're seeing, obviously, is the positive inflection now of, let's call it, wave 2, of now the higher penetration rates on Level 2, Level 2+ technology coming into the fold.

Great. I wanted to ask one, too, about what you have been calling Smart Vehicle Architecture. What does the trend towards Smart Vehicle Architecture portend for Aptiv from both a content per vehicle perspective but also from maybe a market share perspective? Do the technical requirements and innovation required to facilitate that transition towards Smart Vehicle Architecture, does it promise to put some distance maybe between the industry leaders in the field and some of the smaller competitors?

Yes. Yes. So I think kind of like the last one. Let me talk a little bit about what we mean when we say Smart Vehicle Architecture and then -- and why that's so important to us. As we've seen with ADAS, system complexity continues to go up. Electrification systems, infotainment, user experience, there's a lot of content coming to the car. The vast majority of cars today, when you think about their architecture, are hardware-centric architecture. So there's a lot of individual electronic control units, or ECUs, that have software in them, and it all operate somewhat independently. And it's the job of the OEM or a Tier 1 like an Aptiv to try to figure out how to make it all work together. This has been incredibly difficult for the OEMs. And in fact, it's really reached the point where that approach doesn't work anymore. As we go into Level 2+, Level 3 systems, it just completely breaks apart. And so what -- over the past, about 2 years, we've been looking at, well, what needs to change and how should it change. And the solution of that is what we call Smart Vehicle Architecture where you really do a couple of things. The first is you separate that software from the hardware that we're talking about. You decouple those. You get rid of the embedded model. The second thing is you separate I/O from compute. So you basically manage all the I/O in the car, the sensors, the actuation, all of the things along the edge. You manage that separately from your central compute. And that's in a ring -- you can think of that as a ring topology. And this is actually mirroring what's been done in the compute, the personal computing, mobile computing and the server world, but you separate that out. And the reason you do that is for 2 things. One is if you think about the wiring harnesses that are in the cars today, it takes -- for a premium vehicle, it takes 8 to 10 people to actually unpack and install that harness. Think about the complexity of that. We want to break the car into zones and make that much more manageable, and so that it's -- basically, you don't have all of that massive complexity on the interconnect. And so that's another piece of it. And once you've done that, you can then serverize or centralize all the remaining compute on the vehicle in 2, 3 or 4 or however many central compute that the OEM wants. The result of that is a significant cost and complexity reduction for the OEM. And that's the whole point of doing all of that is basically to manage complexity in the vehicle and to reduce costs. Reduce costs, reduce weight, simplify the vehicle assembly, consolidate the whole number of ECUs and reduce cost and weight. And so as we've looked at that, we think by that architectural approach, what we call Smart Vehicle Architecture, you can achieve substantial savings in the vehicle, both in weight and cost vehicle assembly. But also, the vehicle now can be sustained throughout its life cycle with OTA software updates, managing data on the vehicle much more effectively. Aptiv is unique in the sense that we have all of those pieces. We do connectors, sensors, wiring harnesses, compute, so we see it holistically. So as we talk to the OEMs, what's really exciting for me is it's like talking to people that are just desperately looking for a solution and how to get there, and that's what we're providing to them now. We're saying, this is how -- for you, the OEM, here's how you can make that churn. And it's been -- we have 2 advanced development projects now, and we see between today and 2025 a big shift really happening with vehicle architecture. The final point I'd make is electrification is accelerating now. Because when an OEM look at BMW and says I'm going to go with battery electric vehicle platform, it's a clean slate. They can really think about architecture from a fresh point of view, and all of them are smart enough to realize they can't be offering a BEV with a legacy architecture. It has to be Tesla-like or it has to be a digital architecture. So we're seeing a strong pull now in that direction, and electrification has further accelerated that.

Let's check to see if there are any questions in the audience. There are a couple.

[indiscernible]

Yes. So the short answer is yes. Now -- and we do that in a couple of different ways. One is the -- one is organically in terms of developing automotive software engineers, which are different than the software engineers that would go to a Google or go to a Facebook. And so that is an ongoing process that starts with our engagements with universities all around the world, getting that pipeline and then making sure that we have those talent development programs in place. The other way is through acquisition. So when we acquired Ottomatika, which was a CMU spinout, when we acquired nuTonomy out of MIT, the point of that was to get the capability that those individuals had. I mean they had very specific skill sets for doing automated driving, which is a robotics problem you're solving, not a classical controls problem. So it's a very different skill set. And so where we need to acquire talent through acquisition, we have done that. The third avenue is really, as we think about our footprint, establishing offices, what we try to do is establish locations where we know there's a good talent supply, and so Bangalore, India; Krakow, Poland; Boston, there are -- we've been very thoughtful about where we want to locate our offices such that we're in an area that can really attract that talent. So Krakow is now 2,500 software engineers or engineers, basically, the majority of whom are software, has a really rich pipeline. And in that market, I can tell you this, working in the auto space is viewed very attractive. And the final comment I'll make is automotive tech, especially with the advent of automated driving, is now much more attractive. I would -- just to be very transparent. 10 years ago, auto tech really wasn't being viewed as being the place to play. When you think about ADAS systems, when you think about the societal benefit of those systems, when you think about electrification and the impact in the environment, suddenly, it has -- the attributes are there that really attract good talent and leadership. So it's been -- for me, personally, it's been a really enjoyable part of my journey, seeing how this -- how our industry is transforming.

Now last question from Jim in the back.

If you have time.

Yes.

Okay. Are you going to have really kind of 3 regional solutions given the advanced technology that you're doing, China, U.S., Europe? Or you get tremendous leverage across all your capability and doesn't really matter if the heavy regulatory overlay comes into play? And two, with the FCA-PSA merger, we've got 3 or 4, arguably 5 big behemoths that you're going to be servicing. Any thoughts or risk in terms of what they might look to do 3, 4 years from now in terms of internal capability versus what you do today? Is there anything that you guys are worried about on a 3- to 4-year view where the OEM, those big guys might look to do in-house? Does that present any risk to you?

Okay. So for the first question, no. We would -- the fundamental platform is usable in all those regions. Now you're going to have regional variation because of regulatory, market-specific requirements. We manage that today with Singapore, with the U.S., and that's part of what we do already. But the underlying platform is common across all of those regions. You just can't afford to develop 3 individual platforms. To answer your second question, that dynamic of -- the balancing act of what content does the OEM provide versus the supply base is providing, that's been going on forever. And so that's not a new phenomena. I think really the issue here, what you hear mostly about is the OEMs wanting to get more control over the software that's in their vehicle. And that doesn't mean they have to do all the software or make it all or integrate it all, but they do need mechanisms for controlling that software. And actually, one of the key attributes of Smart Vehicle Architecture, what we talk about, is it enables the OEM to get their arms around the software in their car. They can control it. If you do that, then there's a lot less concern or really no concern about, well, who does what. We'll do whatever works best for each OEM, but you have to -- we have to help them get control of that software, which they don't have today.

And it looks like we're out of time, so please join me in thanking Glen and Elena. Thank you so much. Appreciate it.

No. Thank you.