This is Tom Rowland's Reese, and you're listening to Switched on the podcast brought to you by Bloomberg NEF.

Demand for high performance chips in vehicles is taking off as we move towards more automated and connected automobiles with new safety standards and smarter driving systems, meaning that what's inside the car is now just as important as what's under the hood.

As vehicles get more intelligent, the relationship between car makers, chip makers, and suppliers is also changing fast.

What used to be a simple supply chain is now a web of partnerships, with chip companies working directly with automakers and car makers developing their own chips to stay ahead.

And these new processes are doing more than ever before, running both the driving systems and the InCAR software that keeps vehicles connected and aware.

But as chip makers move deeper into vehicle design and automakers start developing their own chips, who really holds the advantage in this new landscape.

On today's show, I'm joined by Andrew Grant, Bloomberg and NIF's head of Intelligent Mobility to discuss his team's note AI chip to turning vehicles into supercomputers.

BNAF clients can find this note, along with AI and automotive research by heading to BNF go on the Bloomberg terminal or visiting BNEF dot com to learn more about how BNF approaches strategy research on the energy transition, including developments in commodity markets, sector trends, and the technology shaping the future.

You can find more information at BNEF dot com.

So let's explore how technology and intelligence are reshaping the global auto industry and what it will take to turn automation into action.

Andrew, Welcome to the podcast today.

Thanks great to be here.

I've been looking forward to this because the intelligence and automation of the automotive sector is such a fascinating topic, and I think it is one of the ones that conjures images of a brave new world, a bright new future with real changes.

Now.

Obviously, as an analyst, you'll typically finding your pouring cold water on people who get excited about this stuff, but I've been looking forward to this just to learn a little bit more about how things are progressing there, and in particular, you know, we're talking about how AI chips and for vehicles, how that space is evolving but let's start with just more broadly, the automotive semiconductor market, Like, how big is that as things stand.

Well, thoma's always been a big market, and it's getting bigger.

As the semiconductor industry as a whole grows, so does the automotive sector.

So semiconductors in twenty twenty four were worth about six hundred and thirty billion dollars globally.

The automotive sector over the last decade is held fairly constant, somewhere between twelve and seventeen percent of that last year has had about eighty billion dollars.

So a lot of money from the automotive psyche gets spent on chips that go in vehicles, but there's quite a wide spectrum of trips that actually make their way under the hood.

You have these domain controllers or kind of simple chips that have a very specific job in the vehicle that control the air bags, they control specific parts of the vehicle.

But what we're seeing more and more are really high performance, really complex chip sets and computers that are being put together and going into these vehicles that are doing really complex sasks, and they're quite expensive chips in themselves, so it's really changing the cost of putting a vehicle together and the balance of parts that go into a vehicle.

That's so interesting.

So we're kind of we're moving from a world where the automotive industry is maybe thirteen percent is pretty significant, but it's still a fraction of the overall global chips market, and the chips are a relatively small fraction of the overall cost of producing a vehicle.

Where in the future we're converging towards chips being a much bigger part of the overall value of manufacturing a vehicle, and chips and vehicles being a much more important of the chip market as well.

It's not two sectors that you naturally in your head put together, chips and cars, but their overlap is getting quite a lot bigger.

Yeah, absolutely, and that's only going to increase as consumers have more demand for automated driving features and advanced safety systems that are in these vehicles, as more robotaxis into the market and are deployed on streets.

Also, I should mention, as consumers have a greater demand for info, attainment and smart features in their vehicles.

I've heard some automakers say that it's not that interesting to be an electric vehicle anymore, you've got to be an intelligent, connected electric vehicle.

So there's a really significant growing demand for these types of computing equipment.

The numbers that you've mentioned, you know, eighty billion dollars of a six hundred and thirty billion dollar market and growing.

That's pretty impressive.

But I mean, I was reading somewhere that suddenly, like the global automotive industry is three point six trillion, and in the next ten years it's going to be reaching something closer to seven trillion approximately, and the two biggest drivers of that is going to be electric vehicles and automation.

So this is a big business in itself, but it's also it's kind of like a lever that's influencing an even bigger business to get even bigger.

So how are all these chips changing the automotive industry?

How is it driving that change?

If it really is.

I mean, certainly a lot of changes of what you mentioned the electrification side and all the challenges that brings to incumbent automakers, and it's introduced a whole bunch of new automakers to the field that are building out their vehicles around an electric drive.

Train, which does tend to be more suitable to some of these computing devices being included in the architecture.

They can build them around an electric motor and a really powerful computer, and it can do a lot more than if you are trying to kind of redesign an existing platform for an internal combustion engine.

So that's changing perhaps who has the advantage in this space, but it's also changing the supply chain quite significantly.

So previously what you might have in terms of the relationship between the semiconductor, the chip maker, and your car maker, there'll be a bunch of different players that sit in the middle.

There, your fabs, your tier one part supply from the automotive side that would work with the semiconductor players, and then they would flow it through this nice orderly supply chain and everyone would take their turn.

And there's kind of a shift taking place where the chip makers need to work more closely with end product suppliers, whether that's the automaker, the car maker, or the robotax you developer, to really make sure that the chip is capable of delivering that end products in the way that the en product supplier wants it to be delivered.

So on moving away from this orderly supply chain into something more that's more like a network where you still have tier one part suppliers doing a lot of integration or tier two part supplies integrating chips into a component that becomes a larger component that ends up in a vehicle.

But you really need a lot more communication and partnerships between the chip makers and the end use robotax or automaker.

So the whole supply chain is changing.

So let's talk about how this is impacting the supply chain.

And I think it's really interesting to reflect on maybe the pattern we see in a way because we've talked about how chips are becoming more central to the future of automation and how that's shaking up the supply chain, and maybe it's not quite as simple as you buy chips from chip supplieres, you put them in your car, and then your car does whatever it needs to and its its application.

It reminds me of how things have evolved and become complicated with cars and lithiumine batteries.

You know, when lithumine batteries it became clear that that was central to the future of transport, and also that transport was central to the future of lithiumine batteries.

Certainly, it wasn't just like you had suppliers that you brought from You had car companies investing in battery manufacturing.

And so I'm wondering, you know, is there a parallel kind of muddying of the waters happening here?

I mean, I think you've already said that there is, But tell us more about this muddying of the waters.

How is this getting more messy?

Who is doing what?

Well?

That's a very interesting analogy in comparison, and I think it holds true in a lot of ways.

So if you start looking at this from the chip maker side, to make a compelling product for an automake or a ROBOTAXI developer that is trying to automate all or some components of driving, you need to basically build out a test case, build out a product that you can prove can do the thing at the end of the day that you are trying to get it to do.

That is a very detailed process, and you end up having to do quite a lot of training of self driving vehicle algorithms.

You need to have a lot of data sets on hand, you need to do a lot of the work that the end customer is going to do anyway, so why then don't you just do the end customer job as well.

So we've seen some interesting examples of chip makers that have developed these prototype self driving systems that they were using to kind of convince customers that their product is great, and now they are starting to actually offer more software attached to these chips and these computers so that they can do more of that self driving at the end of the process.

I think most notably is in Video, who have recently signed a deal with Uber of automakers, whereby in Video is not just going to be supplying computing power and putting that in vehicles, They're actually going to be supplying a lot of the software that does make self driving decisions, and that's a bit of a shift from what we've seen previously.

We've also seen this from Qualcomm, which has a partnership with BMW.

They have a partnership with a few automakers, but they announced a little bit earlier in the air that they would be providing this Advanced Driver Resistance Systems or aid ASS package that is really built on a lot of Qualcomm work and BMW are putting it in their vehicle.

And previously you might have seen that happening through another intermediary, somebody else designing the self driving component of that, or the automated driving components of that, or even the automaker taking on the automated driving components of that.

But the chip makers are not only working more closely with the automakers and the kind of robotaxi suppliers, but in some cases they're taking over more of those responsibilities further down the Valdy chain.

So what you're saying is that a chip maker in the past would have sold chips to a car maker, and whereas now they are actually involved in the software that that's used for and in some of the end applications.

So they're not just they're kind of in a way going beyond even the automaker in terms of how they're thinking.

It's becoming a lot more integrated.

Have I understood it correctly?

Yeah?

I think so.

I mean they are providing not just hardware, they're providing software as well, and that software is doing more and more.

I suppose.

If that is the case, then if you're an automaker or a software provider, then you say, well, why would we let the chip makers capture all this value?

Like, if this is what's happening, then it's anyone's game.

So are we seeing automakers investing in chip design, chip manufacturing?

You know?

Is there sort of some competitive investment going the other way?

Absolutely so, I think most notably you've got Tesla doing this.

It is incredibly expensive to develop these types of computing systems.

Tesla's spending in the region of Timberland in Capex.

A lot of that's going towards building out data centers that can train their self driving vehicle algorithms, but a lot of it is also going into building out their own computing hardware that's going into vehicles.

But this is a very difficult thing to do.

There's a huge skill set from some of these chip suppliers that have been working on this for a very long time, and it's not an easy thing to replace.

It's very expensive to build out these capabilities.

Some other automakers that are looking back and saying, we're not going to let the chip makers be such a huge component of our value chain.

Companies like xpaying and Neo, both companies founded out of China, who will be wanting to do this for similar reasons to Tesla.

They want to own their own destiny.

But at the same time, Chinese companies have a bit more of an incentive from a different angle to do this because it is more difficult to get some of the leading IT chips if you are building vehicles in China and operating in China because of the very difficult trade environment that we find ourselves in at the moment.

Okay, So that's interesting.

So this question of whether or not to make your own chips is different for the Chinese automotive industry than it is for the automotive industry elsewhere.

Yeah, there's certainly thing other things to consider here.

So then why is Tesla going that way?

Tesla has tried to be as vertically integrated as possible across pretty much everything that it does.

It's wanted to build its own hardware.

Its early rollout of its robotaxi system is it's very in house approach.

They are not working with many, if any partners on that front.

They want to keep as much of the value construct inside the company as possible.

They also have some ambitions to use the technology that they developing on the chip front in other areas.

Whether that's kind of the Optimus robots or other projects that they're putting together.

But it's a bit that they will be able to build these chips that are best suited for the purpose that they are using them, rather than getting some other company in and trying to kind of either adapt a product that already exists or that somebody else is designing for their self driving service, or they will have to build their own system that kind of does the thing that they needed to do.

I mean, doesn't also maker have any competitive advantage in this situation.

And I realized it might be different for Tesla because obviously their owner also has investments in the tech industry, so maybe there's more synergy there.

And I don't know, I'm just kind of speculating here.

I mean, I wouldn't say an automaker has a particular advantage at building the types of chips that we are talking about here.

We're talking about the real bleeding age.

I mean, these chips are being produced at four nanometers at the moment, So for some of the most cutting edge chips that are used for self driving vehicle development, that's a fraction of a human here.

They are very few chips globally that are produced at anything more advanced than that.

Just having that fabrication capability, there's kind of one or two suppliers globally that are able to do that.

So really use still need to work with the existing semiconductive value chain to get these things built out.

So I wouldn't say that an automaker has necessarily the capability to go and alone, and even somebody like Tastler's working with Samsung Foundry to actually build out the chips that they design.

And I suppose maybe to some extent, it depends on you know, you're making a bet on in the future to what extent are automotive applications the main driver of AI chips in the future or not, Because if they're not, then you really don't have a competitive advantage.

But if they are, maybe you do have.

Maybe this is a logical pathway.

I don't know, it's super interesting and I think you touched a little bit on h We've talked about chip makers expanding in the value chain, certain automakers, particularly in China, getting into the chip side of things.

I think you touched a bit on partnerships before.

But you know that seems like the third way, So could you just can we dive a bit more into like what that strategy looks like.

Yeah, so we've seen lots of partnerships sprung up between chip makers, automakers and robotaxi suppliers and a lot of deals have been signed.

So Mobile is a company that's gone around and made numerous partnerships.

We track it in the region of at least fifty companies or automotive brands that they have agreements to put their chips in those vehicles.

But really on the robotaxi side is wherein Video has had some strength.

They are in the majority of different robotaxi developer brands that are actively testing or reaching advanced or operational stages at the moment.

So really Video is kind of the chip of choice for the robotaxi supplier Mobile Eye it has the most partnerships.

But when you look at who's collecting the most revenue from this space at the moment, that seems like Qualcomm, who have been racking up a decent number of partnerships that also have kind of revenue figures from the automotive space to back that up, although I should clarify that with in Video's automotive exposure not just being tied to self driving chipsets like the Mobile I and then video ones.

They are also they have a number of entertainment products that are part of the story.

Just as an example, a company like in Video, how much of their current revenue is for chips for the automotive sector.

Well, in Video on track to make about two billion dollars this year from the automotive sector, which sounds like a big number, but point of Video, that's just one point three percent of their total revenue.

Okay, I mean it's a big number.

From the notes I've got, I'm seeing Mercedes Benz within Video, BMW with Qualcom, Volkswagen with Horizon Robotics, Porsche with Mobile Eye, and just an observation, these are all European companies.

Is that a trend, that is European automakers seem to be pursuing a strategy of partnership or is that just the examples that I happen to have seem to paint that picture.

I think that's the examples that you have that are painting that picture.

You could say the same thing about your GMS and forwards also going that route.

Really you don't have that many automakers that are looking to go alone here, it's really just tied to a few, and they all tend to be new aid EV startups that are building their vehicles around this idea that they will be electric and they will be intelligent, and they've got to handle the electric drivetrain, in the battery and all of that.

But then they also have to build out a smart infrastructure that can be compelling to consumers.

Something that we've noticed quite recently is if you look at the number one factor that consumers in China instead they look at when considering purchasing a new vehicle, the intelligent components are on number one on the servers that we've seen.

It's not isn't an EV as it clean, doesn't have affordable mileage or any of those other features.

It's the digital aspects that come from having these advanced chip sets in the vehicle.

I mean, we're talking about two industries, cars or automotive more broadly, automotive and AI slash chips, both of which, in very different ways governments tend to think of as strategically important.

Whether it's from a job's perspective, whether it's from a security perspective, so I imagine that policy is playing a role here as well.

There's obviously the aspect of safety and quality standards and you know regulation more generally, which must be an aspect of it.

But yeah, what are different governments doing in respect to this kind of this evolution we're seeing and what's happening around chips and vehicles.

Yeah, I think about this in two ways.

So one there's the regulatory side that putting safety standards in place, making sure these are built properly using the proper procedures and kind of development protocols.

And really these are chips that are taking on a greater and greater responsibility within the vehicle, so they do need to be automotive grade and everything that that means, not just on the physical hardware side that they can withstand shocks and bumps, but also that the software is developed using the right protocols, that there are cybersecurity measures taken that make sense.

So those restrictions make a lot of sense to me, and there's a growing number of in there.

But what it really means is it can make it more difficult and slow down the process for developing these chips and getting them into vehicles.

But it's something that I think is necessary.

But then the other role that regulation is playing is how governments and regulators are requiring more safety features and smart safety features that are digitally enabled in new vehicles.

So things like advanced emergency breaking, lane keeping assistance, pediestrian awareness, cyclist awareness within vehicles are becoming not just added features that automakers might put in the vehicle to make their vehicle more compelling to consumers, it's something that's actually starting to be mandated in more and more places, and that only drives the demand for these types of chips up.

That's interesting.

I remember also, I mean, apart from the safety aspect, in a previous life, whilst still being at BNEF, I used to cover energy efficiency and at one point in time, LED lighting, believe it or not, was the big new thing.

I know that ledlighting.

We take it for so for granted so much, And I remember being in a lighting conference and someone explaining to me how the cutting edge application for LED chips is in vehicle headlights, and the reason for that is because the standard of quality has to be so much higher and the fail rate is so much lower because it's so much more expensive to have to change one of these things out than it is changing a light bulb in your home by comparison.

So I mean, I suppose a question I have here is obviously from a safety point of view, the bar is already pretty high on these, But I also wonder it's like in that point of view of vehicles are not things that you easily just replace parts on or take the part.

You know, everything is very designed as a package.

Does this mean this is also raising the bar for chip makers, like this is the real If you can do it right in vehicles, you can do it right anywhere.

Yeah, exactly, So I think that holds the stakes are very high, so it's important to get this right.

Part of it that it's also a market that interacts closely with a lot of consumers, just like your mobile phones, sor it does cars, arc and products that people want to build out a product and have their relationship with the consumer.

So that's a big part of it as well.

I want to ask you about, you know, now we've spoken about everything, what you think the future looks like, both in terms of the developments we can expect in the next few years, but also how important vehicles are now to the development of AI, because I think that when we think about AI, everyone's thinking about AI generated content because that's front and center.

But vehicles could be the application that really drives this.

Maybe we'll be interested to get your view.

And then how important and how much of a difference does this make to how cars and the mortus motive sector generally is going to evolve.

My colleague Shung who who is one of the authors of this note that we're discussing, he likes to say that self driving vehicles are the most tangible way that humans will be interacting with AI for the next few years.

So this really is something that we're going to interact with the lot and it's going to be very close to us.

And then what we have is an indutry where the main components of the vehicle are really starting to change, and the bill of materials, the cost of all the different components that go into those vehicles that's being reshaped.

Suddenly you have an industry that has the battery of the electric vehicle accounting for a large chunk of fit a third or baut a third of the vehicle.

And then you have this growing share that's going into the computing hardware and it's really a big supply chain shift to make it that's trying to build a vehicle, or when you've robotaxi developer that's trying to build a robotaxi.

It's so interesting because I've been thinking about, you know, we're in this period where having batteries is a really important thing if you're an automaker that it wants to jump on the electric vehicle wave.

But I've always asked myself the question, you know, if that's really going to get kind of commodified, if that's the right word, And you know, also the cost is going to go down and the margins on that aspect of the value chain might become really thin, whereas in the internal combustion engine vehicle world, actually because they're so complex, actually there's a lot of margin and differentiation in there.

What is going to be the thing that separates out a really nice car from a really basic car and from everything you're telling me that might well be everything we're talking about today is like how clever is your car?

Becomes the question because your car will already be really fast because it's electric, It will already be really economic to run because it's electric.

So what is there left, and maybe it is this question of how clever is it?

Yeah, how much of the driving can it do for you?

What are the kind of infotainment or entertainment options that you have within the vehicle?

How well does it connect to the rest of your digital life.

That will all be incredibly important, and automakers and kind of everybody throughout the value chain of looking at that and thinking how much of that are things that we can actually charge for and make money from.

Can we sell subscriptions to some of these self driving features or do we need to kind of bake that into the cost of selling the vehicle because consumers are just going to expect it and won't want to pay extra for it.

So these are kind of a lot of the questions that the automakers are looking at at the moment.

Really fascinating stuff.

Thank you, Andrew.

This has been so interesting.

I sometimes just wish I had a time machine.

Firstly, I could get rich with that, but secondly because the possibilities here are so exciting and transformative, and I just would love to know what the world looks like in ten years time, what the world looks like in twenty years time, what the world looks like in fifty years time with everything we're discussing.

So thank you so much for getting me excited.

I really appreciate having you on today.

Thank you very much for having me.

Today's episode of Switched On was produced by Cam Gray with production assistants from Kamala Shelling.

Bloomberg NEF is a service provided by Bloomberg Finance LP and it's affiliates.

This recording does not constitute, nor should it be construed, as investment advice, investment recommendations, or a recommendation as to an int or other strategy.

Bloomberg ANIAF should not be considered as information sufficient upon which to base an investment decision.

Neither Bloomberg Finance LP nor any of its affiliates makes any representation or warranty as to the accuracy or completeness of the information contained in this recording, and any liability as a result of this recording is expressly disclaimed.