The Garage by Sonatus

Today in The Garage,

we're recording live from the
COVESA all-member meeting in

Berlin, Germany with
Ward's Intelligence.

Let's go!

Welcome to The Garage.

I'm John Heinlein, Chief
Marketing Officer with Sonatus.

We're recording live at the
COVESA all-member meeting in

Berlin, Germany.

My guest today is Maitê Bezerra,

senior principal analyst for
software defined vehicle for

Wards Intelligence, which
is now part of Omdia.

Maitê, welcome to The Garage.

Thank you for having me.

Nice to see you.

And, you know, we always start by
getting to know our guests a little bit,

so I wanna give you a chance
to tell us about you and your

background for a minute.

So, John, I am an analyst looking
at software defined vehicles.

So my life, no my job,

is actually trying to understand
this transition in the

automotive industry.

But before looking at
software defined vehicles specifically,

I looked at all the
automotive tech.

And before that,
I was an academic.

We're thrilled to have you.
Thank you for joining us again.

And usually, we ask our guests
to tell us a fun fact about them,

but I have a special fun fact
to share on both of our behalfs.

You have the privilege of being
the first guest to be on The

Garage podcast for
the second time.

Ah, that's brilliant.

You're the first one.

And so we have a special surprise
for you, which is, we have,

present it's the the Garage
podcast mug that we'll give to

all of our guests,

but you're the first recipient
as our second it's our first

ever second time guest.

Oh my god. That is
an honor. I love it.

Thank you for being here.

But we're we're so plea pleased
to to talk with you today.

And you and I just came off
stage here at the COVESA

all-member meeting where we're
talking about the annual SDV

survey that you carried out,

and this year was
sponsored by Sonatus.

And, we wanted to share a kind
of a preview of those conclusions

today to our podcast listeners.

SDV is such a, as we
said, broad topic.

There's so many parts to it,

and I I think you and I both
learned a lot with this survey.

So can you introduce the survey?

You've been doing it
for several years.

And talk about what the goals
and what we what we did.

Sure.

So the survey in itself,

it's something that Warts
Intelligence has been doing for years.

Because the idea is really to
understand the transition that

the automotive
industry is going,

but also how things
are changing.

Right?

Because we know this is
an incremental change.

So it's really interesting to
see the new trends and the new

aspects that we address and
that we find every year.

So this survey is,

usually a survey that we
deployed in North America,

some countries in Europe.

But now with the support
of Sonatus, I'm really,

really happy about this.

We were able to expand
it to seven countries.

So now we have, US, Canada, UK,
France, Germany, Japan, and China.

And this is an industry survey.

So we deployed the
survey to automakers,

to to tier one suppliers,
tier two suppliers,

including software
and chipset as well,

consultants, and
trade associations.

But all related to automotive,

and most of our
respondents are working,

in some parts of the vehicle.

Yeah.

Fully seventy five percent of
the respondents were right in

the automotive ecosystem,
not counting those that are,

you know, consultants or
may have a broader role.

So really and it's not a
consumer survey either.

Really, people right in the
thick of SDV and and doing the

innovation, so that's fantastic.

And we also we're we're so
excited to have that many

countries because it's
such a worldwide effect.

And one of the
things we learned,

and I'm sure we're gonna talk
about and both of us have our

iPads because there's so many
data points in the survey.

We found so many regional differences
in the in the respondents.

So I think we can talk about
that a little bit as well.

Let's do it.

Alright. So let's jump in.

Earlier at the COVESA
all-member meeting,

we were on a panel talking
about artificial intelligence.

And so let's start by talking about
some of the conclusions from AI.

AI is has become more and
more commonplace in vehicles.

And yet, as we mentioned on
the pan as the panel earlier,

what we're finding and my
observation is still it's only

very early days of AI.

And what we were curious to
find in the survey was what and

the question we asked was,

what are the most promising
use cases of AI in vehicles?

And you might think that things
like chatbots and, you know,

navigation, things like that,

entertainment system
would be the top.

Actually, those were
the bottom of the list.

The top applications of AI
in vehicles that are the most

promising were things that
are largely not done yet.

The top one was
smart diagnostics.

So exciting to see that.

Another one was dynamic vehicle
performance calibration,

things like
ongoing optimization, range

optimization, vehicle driving
and comfort personalization.

And number five number four
was routing and navigation,

so that's kind of normal.

And number five was
predictive maintenance.

So fully four of the top
five applications of AI are things

that I would say are broadly
not done at all yet in vehicles

in any kind of
widespread application.

Was that surprising to you?

Definitely.

And also, I think one of the things
that is surprising is to say,

we often think about the
infotainment when we think about AI.

Most of the the applications
that we see over here about the

demos are in the infotainment.

But I think what we what this
is demonstrating is that to to

create value in the car,

you need to focus on things
that are car centric,

that cannot be replicated
with your phone,

that will not be
competing with your phone.

Right?

So this would drive value to the
customer when they are driving the car.

Right.

And, you know, there's
a big you know,

a lot of complexity in
terms of is it CarPlay?

Is it Android? Is it in
integrated in your vehicle?

But that's all, as
you say, infotainment.

How can we do something that's
much more integrated in the car?

And the survey plays that out
that there's a desire to really

create value in the car.

And one thing that I I find
interesting is that if you look

at these, use cases, so
smart diagnosis, for example,

or dynamic vehicle
performance calibration,

these are not simple
AI applications.

You actually need,

a next generation
EE architecture,

where you have your sensors
connected to the domains,

and you also need edge AI
models that are capable of

deriving insights about the
user, about the vehicle,

the environment, the cloud,
and then take action.

That's right.

So, I thought it was really
interesting because we're

talking about complex AI
cases that need AI, the edge,

that need flexibility
in an SDV architecture.

Yeah. We started with this
AI. It's such a hot topic.

We started with this question,

but I think it's so
illustrative of some of the

trends, the broader trends
in the in the thing.

You and I met at CES this
year, and we showed off,

some more AI applications for
with something we called AI

Technician Builder.

And it really shows off this
exact idea of how do you link AI,

and insight and into
vehicle data in a way that's

fundamentally more
value creating.

So I think this this plays out that I
think we're on the right track with that.

That's brilliant.

Let's go into some more things.

One of the other,
questions we asked was,

what's the best way to
build customer loyalty and

potentially revenue with SDVs?

What were some of the
trends you saw from that?

I think this is interesting
because if you look at what the

industry think will lead
to loyalty and revenues,

it's automated driving.

And this is the top application
in nearly all regions.

And I say nearly because, it
doesn't appear in Germany,

not even among the top three.

And in China, it's second.

And I think China was really
interesting because what they

think would, drive value and
revenues is their AI assistant.

And just to, clarify, the AI
assistant they're talking about

is not the AI assistant that
we are used to having in our

phones or in our
smart home devices.

This is an AI tool that
is capable of, again,

communicating with the vehicle
functions Controlling more of the car.

Controlling the car.

It's something that goes way
beyond the infotainment. Yeah.

It's something that will
have access to your ADAS and

autonomous driving systems.

The car functionalities of the
vehicle will will enable you to

adapt the vehicle
ambient, for example.

So it's, it's a very robust
application, actually.

You need very robust AI
application to achieve that,

which reflects the
state where China is.

That's right. And then
you mentioned, ADAS was there.

And then as you got
into other regions,

then it became more complicated.

Things in Germany,

they they feel that the driving
experience is important.

Things like personalization,

vehicle ride customization
was important in Germany,

also in the UK.

Personalization actually
applied in in most regions as

well, making the vehicle
your own, which is, again,

not just the infotainment,

but how can I make it that it
feels differently, you know,

whether it's your driving
settings or performance or

climate or many things like
that, really personalizing it?

And things like, you know,
even downloadable themes,

we were saying in the US,

peep people seeing in North
America seem to really care

about personalizing
their vehicle,

whether it's lighting or themes
or maybe their favorite sports

teams or whatever.

Great.

And I think, one important
conclusion to drive here is we

are not gonna reach a point in
which a single application will

be the most valuable everywhere.

Every region has
different preferences.

And I think the important point
for the OEM is first to find

out what these preferences are,

but also have the flexibility in
their vehicle platform and, of course,

in their organizational
processes should be able to

address these
different preferences.

That's a great conclusion.

And, you know, Sonatus,

as we're saying and our big
theme this year is that we feel

SDVs are a platform
for innovation.

And it's exactly that
idea. It's not one thing.

It's not make the IVI smarter.

It's not just make ADAS smarter,

but it's a range of
different things.

And even as we've said
before, we don't always know,

when a new application
is gonna come out.

We need the infrastructure of
the car to be flexible enough

to adapt to something that
comes tomorrow that is

compelling, and we don't
obsolete our vehicles,

and we don't fundamentally
create disadvantages.

Precisely. That's great.

Another thing we asked
about was vehicle data,

and we asked customers
what were the highest value

applications for vehicle data.

The the top answer we
saw across almost across most

regions was, vehicle
data monetization.

And and that might make sense,

and I think there's
there's a continuous,

journey in the industry
to say, how do we provide,

value creation, perhaps
revenue services?

You just finished a panel a minute
ago looking into that as well.

But what was also interesting
was right below there,

the answers differed
a lot regionally.

And I think that that suggests
that there's no single way

people are looking
to monetize data,

but rather they're looking for a range
of things which may vary regionally.

What were some of
the trends you saw?

Well, as you say,
they're looking for monetization.

So I I would like to
link this to the main challenges because

this was another question
that we had in the survey.

So what are the main challenges you
were finding through to scale SDVs?

And, if you look at it, by far,

it's justifying
return of investment.

It's justifying the cost.

It's optimizing costs, and
finding business cases.

So, it's not
surprising, I think,

that what all OEMs want to do
is to monetize the vehicle data.

Right?

But as you say, and we heard
this in the discussions,

at COVESA now, is it still
difficult to find out,

how to make money?

We we still haven't found a way
to actually make the vehicles

extremely profitable.

The OEMs are still struggling
to find what will be this

application and
the means for it.

Again, it's it's something that
having the architecture and having the

flexibility and
understanding the customers,

they will find in the future.

In terms of the
regional preferences,

I think one thing that is
interesting is in China

that they say that data
driven product development and

improvements is one of the most,

high value vehicle
data applications.

Kind of a feedback
loop of improvement.

Exactly, exactly.

And it shows that the other
regions are still a bit behind

in terms of this.

But China really understands
that you need to understand the

vehicle in your customers
to evolve your product.

That's right. Very exciting
and very interesting.

Another one that I thought
was interesting is energy

consumption and
efficiency optimization.

We mentioned a minute ago in
terms of AI that this that one

area of feedback about
tuning the vehicle,

I know a number of OEMs have
said they've challenged their

teams to provide ongoing
optimizations over time so that

especially battery electric
vehicles particularly can

improve over time as maybe the
battery fades a little bit over

lifetime so that they can
maintain a common experience

for their customers.

So you're seeing that in
a number of regions was

interested in ongoing
energy optimization.

I like this.

I I think it's very, very
important, for the cars.

Because today, one of
the problems we have,

even with navigation.

Right?

Because navigation
is, as we saw,

one of the areas that
we're going to see value,

in in the vehicle.

But if you don't have access
to real time vehicle data,

if you don't have access to,

the state of the driver and
the state of the passengers,

how many drive how many
passengers you have in the car,

and you don't have,
you know, cloud data,

to understand the the road conditions
and what they're going to face.

Your estimation of range is
not going to be accurate.

And this, of course, would
lead to a bad user experience.

So, of course, SDV has a
lot to offer in this area.

That's great.

I mean, another example
is is battery algorithms.

The, you know, battery
management systems, you know,

while we think of that
as a solved technology,

the reality is that chemistry
and the technology is evolving

so fast as we're
learning more things,

and I fully expect those
algorithms are going to change.

We want to make sure vehicles
are upgradable for many

reasons, not just the the
car the person-facing things,

but also the
infrastructure inside.

So then we talked a lot about,

a lot of infrastructural things,

and and there were so many
questions in the survey.

We can't get to them all.

But this was one that stood
out to both of us, open source.

What did you find?

And also start by telling us how this
year's answer compared to last year.

Well, I love this one,

because I think open source
is really going to bring,

you know, a lot
of optimizations,

to to the industry.

So last year, we asked,

what the industry thought
about using Linux,

in safety domains, okay, to
make it clear, safety domains.

And the answer was twelve
percent of the respondents...

Last year?

Last year.

Twelve percent of
the respondents were,

would consider that.

Now we are asking, in 2030,

what do you think would be the
leading OS in safety systems?

Leading OS in two
thousand and thirty.

Leading OS. Fourth
position, Linux.

And thirty eight percent of OEMs
think Linux will be used for safety.

I mean, this is
a drastic change.

I think it has been, of course,

motivated by the announcements of
companies such as Elektrobit and Red Hat.

They're putting a lot of effort in
bringing, Linux into, safety domains.

But the industry
is clearly ready.

It seems that they do have the
appetite for open source now.

Yeah. Linux, Linux is
such an incredible story.

I I I think I told you that.

I I worked with Linus Torvalds many
years ago in my previous company.

And so to see at that time when
it got adopted into enterprise

applications, which was
revolutionary at that time.

But who thought that you could trust
Linux with enterprise applications?

And I think it's it's growing,

and it's only continuing to mature
even into safety applications.

Still a lot more to go, but
but great progress, I think,

is exciting to see.

Yes.

And, you know, just sorry,

because I still think
this is an amazing story.

It's brilliant.

Because, when the
survey was deployed,

we we still didn't have, announcements
saying Linux, it's certified.

So this shows an appetite
before we even have full

solutions out there.

I mean, we do have
a few solutions,

but not necessarily just
certified the Linux in itself.

So it's, I think
this is amazing.

Very exciting.

Yes.

Very exciting.

Another thing we we looked at,

was the time horizon of
deployment of various

technologies, and there's many,

many slides here
which we'll share.

We'll talk about that in a
little bit of how people can

get more access to that.

But let's let's look at
some of the trends we saw.

And the first one is AI.

We talked earlier about
the enthusiasm about AI,

albeit in the future, and
the data played that out.

We saw that, AI based
applications for ADAS and AV,

which, of course, you know,
there's some amount of,

autonomous driving and some
amount of driver assistance in

many cars today,
certainly many new cars.

You saw that,

the expectation is that the
predominant vehicles will be

deployed by twenty six, 2026,

2027 will have AI
deployed in ADAS, AV,

and that's probably
not surprising.

But what we found was in
areas not ADAS and AV,

the horizon was much later,

more like 2028-2029 most
regions felt like AI would be

deployed in non ADAS things.

And then we were talking
about this earlier.

In in parts of the
organization, not the vehicle,

but in AI development inside
the OEMs and inside the tier

ones, more,

pervasive deployment of AI
was much further out even into

2030-2031 before they really
expected AI processes.

Tell me about your your
observations on that.

Well, I honestly I get
a bit worried about these results.

I mean, I think this is a clear
reflection of the industry

today, quite frankly, looking
at how the industry is moving.

But, seeing, the AI in
organizational processes,

something that is going
to be, deployed so late.

And even when we asked in
our question on how do you define

an SDV, right, the
enabling capabilities,

or AI organizational processes
was really at the bottom.

And this worries me because
when we we talk about the main

obstacles to SDV adoption,
what we hear now is that it's

the organizational processes,
it's the culture, it's the silos.

Enough talents to handle it.

Exactly. And not
only enough talents.

I think one of the things
we even found this,

to be a problem, especially
in Germany, in our survey,

is that they can bring talent,

but they can't retain the
talent because they don't have

tool chains for
these people to work.

They don't have the environment
for these people to work.

So there's no talent
retention as well.

We saw we saw in the past
years a lot of talent coming from the

big tech to the OEMs,

and we also saw a lot
of them leaving as well.

That's right. That's right.

So I think it's a
it's an opportunity,

but it's also a concern that,

the automotive landscape
has a ways to go.

Yes. To really make
AI pervasive across them.

What are some other,
trends you saw?

Well, then we have the the business
and organizational innovation.

And I think this is, one that
is very interesting because,

when we look at,

the the capabilities that
will be deployed earliest,

we have feedback loop.

But then we have a lot of
OEMs saying not only OEMs,

the whole industry is saying
that this will be deployed by

2027.

Whereas in China,

most of the respondents says
that it's already deployed.

Right.

So I think this this shows
again why they are so advanced.

Right?

Right.

And after that, we have
new business models,

something that is expected
to be deployed around '27.

My honest opinion, I think,

the OEMs, when they think
about business models,

they still have
somewhat narrow view.

Thank you.

Yes. It's like, oh, we're
going to charge subscriptions.

Whereas I was just having a
conversation with an OEM right now,

and we need to start talking
about total cost of ownership.

Right.

We don't hear people talking
about the costs of SDVs after

they are on the roads.

Right.

Because they're used to having
the cost to manufacture the car

and sell this to the
customers, and it's over.

Right.

But now you have cybersecurity.

Right.

Now you need to
maintain your car safe.

You need to maintain
your car secure,

and you need to maintain
your, user experience fresh.

So that costs.

And and, you know, we
we've talked about how,

a lot of people think that
vehicles are smartphone on wheels,

and and we I think you and I agree
that that's a terrible analogy.

But one of the things I observe
about the smartphone market is

the device itself is a small portion
of the value of the experience.

It's the additional
services that get added.

It's the upgradeability.

It's the app store, which has
created a massive, value creation.

And I think the OEMs are
still looking for that,

and I think as an industry,

we're trying to provide
the infrastructure

that lets them do that.

And and there's great progress,
and you're hearing a lot of,

great examples here at the
panels in in this session,

but there's still ways to go.

Yes. Definitely. But
there is progress. Yeah.

And this this is
a a major point.

I think after that,

we see the integration
with the IoT and ecosystem,

the IoT ecosystem
and smart cities.

And It's very late.

Very late.

Yes.

Yeah.

I mean, smart cities is such a is such a
great it's like a windmill on the hill.

Everyone is saying
that it'll come.

But in many, many geographies,
it's very challenging. Yeah.

Even even in China where Asia
Pacific where you might have

thought it would be
coming a lot sooner,

and generally speaking,

a lot of those kinds of
technologies come sooner.

Even there, it was
seen as further out.

Yes. That that's true.

And also the the collaborative
organization and structures,

another thing that with
the the survey results

show that the OEMs
will start deploying by

'28, 2028 onwards,

that should start sooner.

Let's talk about OTA. That
was another huge trend we saw.

All all different parts of
OTA and software updates,

which is so important.

What are some of the things
that stood out to you?

So this is really interesting.

So if you compare
it to last year,

the the respondents said
that OTA updates will be,

would be implemented
between 2025-2027.

Now most of the respondents
say it's already implemented.

Right?

So this is a major
shift, which, yes,

we we are seeing the OEMs
advancing on this, pretty quickly.

But I think even
though that's true,

and we've talked
with this many times,

is the observation is that
even though there is OTA in

vehicles, it's often only
updating a portion of the vehicle.

Yes.

So the opportunity is still there
to update more parts of the vehicle,

than than is
typically done today.

So I think they would
almost be maybe next year,

we'll ask another question
about which portion of your

vehicle is upgradable.

Definitely. Yes. But but then
we have the OTA upgrades Yeah.

Which is a question that we
asked as well. And sorry.

I have to look at my tablet.

It's

says twenty twenty seven.

Right?

So we see the OEMs, looking
to deploy the OTA updates.

And then, of course,

we should ask which part of
which portion of the the vehicle.

I think it's more infotainment,
but it's still yeah.

We are seeing we are seeing
advancements in there.

And then one we
also asked about,

things like virtual development,

and, and digital twins.

And I was surprised to find
that the horizon for that was a

little later than I expected.

And I think virtual development
is a huge opportunity.

A lot of organizations,
Elektrobit is focusing on that,

many others are
focusing on that.

It's still very
much in front of us.

And but everyone realizes that
can has the potential to speed

design cycles and
speed up the timeline,

but still in front of us.

Yes.

Well,

it might be good news.

Last year, they said 2030
This year, we expect [2029].

So maybe we're gonna
see the the timeline.

I think that's a
really great point.

I think if we pull back
and look at a macro level,

all of the trends here suggest,
an acceleration of SDV.

It suggests a
realization that SDV,

which was perhaps
buzzwordy a few years ago,

there's some real, concept,
there's some real value there,

across the regions.

So I think that's the one
huge positive takeaway.

I completely agree.

And it's it's interesting
because there was only one one

technology that we saw
the timeline increasing,

and that was,
interestingly enough,

centralised vehicle
data management.

And that's funny because I
think it's that there was a

perception that was much
simpler than it actually is.

I think there is now the
realisation that what they

thought was centralised
is actually very siloed.

Right.

That the departments that
really needs the the data don't

really have access to it, so
it's gonna take longer to.

Yeah.

We're we're here at COVESA.

One of the key focuses of
COVESA is, vehicle data.

They have a specification
called VSS that we we

participate in and others.

That has the potential to
help some of those problems,

and help to standardize, but,

it's still still
some ways to go.

Yeah. Definitely.

Well, Maitê, we could talk for
hours and hours about this.

Let's tell our our viewers, how
can they hear more about this?

What are some of the other
things coming up where they can

learn more about the survey?

Well, the results of the survey
are going to be in a white paper

that will be published very
soon, and you can access.

We are also going to
discuss these results at AutoTech in

Detroit, and we have
The Garage, of course.

That's true. We'll be
recording there as well.

So I look forward to
seeing you in Detroit, and,

look for this white paper soon.

I know you'll have
it on your site.

You can come to the Sonatus
site as well to find it.

And there's just so
many interesting graphs.

We've just given you just
barely the surface of this.

We we learned a lot from it,

and I'm sure our guests
will find a lot as well.

Thank you so much
for joining us again.

Thank you, and thank
you for my gift.

You're welcome. You're welcome.

Thank you for
joining us as well.

If you like what you're
seeing in The Garage,

please like and subscribe so you
can see more episodes like this.

And I wanna also give a special
thanks again to COVESA for letting

[us] film at the COVESA
all-member meeting here in

Berlin.

We look forward
forward to seeing

you on another episode of
The Garage again very soon.