The Garage by Sonatus

Today in Garage, we're
recording live from AutoTech

Detroit. And our guest today is
Suat Kusefoglu from Capgemini.

Capgemini is a major worldwide
system integrator and a key

leader in the automotive space.

In today's episode,

we talk about the importance
of the evolving system

architecture as
SDVs get deployed.

We talk about learnings from
the Chinese automakers and what

worldwide automakers
can learn from that,

and also the importance of
maintaining a balance between

differentiation and standardization
and vehicle architecture.

It's a wide-ranging
conversation.

I'm sure you'll enjoy.

Let's go!

Welcome to The Garage.

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

We're recording a live
at AutoTech Detroit.

You can hear the noise from
the conference all around us,

but we have so many exciting
guests here at the show.

And our guest in this episode
is Suat Kusefoglu from

Capgemini.

Suat, welcome to the garage.

Thank you for having me.

Tell us a little bit about
you and your background.

Yep.

So I've, worn multiple
hats over the years.

I have roughly twenty years
experience, working at OEMs,

also tool vendors and, and
engineering services providers.

My foundation is
in EE architecture,

systems engineering, software
architecture as well.

And I've recently moved more
into leadership types of roles,

more advisory on the shift
to software defined vehicles,

trying to help find the
right balance in in cost complexity

and finding the right
solutions with partners, like yourselves.

Fantastic. Where are
you based out of?

I'm based in, Novi, Michigan.

Okay. Oh, local so
you're a local boy.

Oh, that's fantastic.

I've been this from here.

That's right. So you gotta
tell us a fun fact about you.

We like to get to
know our guests.

Yeah. So slightly,
nerdy fact about myself.

So I have this
obsession of memorizing,

you know, album titles and,

and and band members
and record labels and,

and such.

Yeah.

It's kinda like my my brain
decided on what kind of

metadata it wanted to, store.

And, yeah,

quite quite So you so you just
love to learn about different

different bands and
the let's see that.

What can I say?

I always like to have a fun
fact to match our guests.

I'm a big language buff. I'm
always learning languages.

So I'm constantly doing, like,

flashcards of
different languages.

My current language of
focus is Spanish. Wow.

And I speak French pretty well
and a few other languages.

But, currently,
I'm doing Spanish.

So that's my there's my
fun fact fact to you.

That's great.

So tell us about Capgemini,

both in general for our listeners
who may not know your company.

It's a very successful
company. Yeah.

And also, both what you do for them
and and what do they do in this space.

Yep.

So Capgemini is a large
organization working across,

many different sectors.

It's roughly, it's over three
hundred thousand employees globally.

We work with clients

both on-site and and
near shore models.

I work specifically in
the automotive sector

focusing on on the
software defined vehicle,

working across the entire stack,

from architecture design to
software integration and then

also helping clients
find the right partners,

for specific solutions.

That's great. And, it's a
huge company, Capgemini.

I've worked with them for years,

incredibly smart
folks that are there.

What kinds of projects
do you get involved in?

So I I focus specifically on,

cross functional
and cross domain,

activities trying
to help really,

find innovative ways
to transition from,

distributed architectures to domain
and zonal, also from,

signal based systems
to service oriented.

Right?

It's a very complex,
task to achieve. Right?

There's a lot of,

moving parts.

Right?

And, getting right getting it
right is is really a challenge.

That's right.

We talk about so many of those
topics here in the Garage

podcast, software, the
evolution of vehicle

architecture, and so on.

So maybe for our
first topic, you know,

you and I both have a
background in electrical

engineering and
in system design.

Tell us about how
you're seeing SDV,

the architecture of SDV is
changing and some of the both

the challenges and the
benefits of those changes.

Yeah so we're seeing a
fundamental shift in how things

are designed right so it's
moving from a discrete ECU to

dedicated purpose tightly
coupled software and hardware

solutions provided by tier OEMs and
and then integrated by those OEMs,

to more, to less boxes in
the in the in the vehicle.

Right?

More complexity and managing,
tier ones that might provide,

off the shelf
software solutions.

So the challenge is really,

managing all of the
complexity around,

contractual, agreements, right,
between different companies,

and then also finding a right
the right balance of how to

find the right software
solutions, right,

and integrate all of those.

So we're moving from, a
finite set of tier ones that

need to be managed, right,

to maybe hundred
and a hundred plus,

suppliers providing
small bits of

accelerator software or

full solutions that need to
be integrated into a box that

comes from a different
type of tier one.

That's a really interesting shift
because as you as you

pointed out initially
and and classically,

there was a hardware box,

there was a software
load that went with it,

they were put together at at,

manufacturing time and
many cases never updated.

That's changing a lot now.

You have a a fewer boxes with
workloads side by side like

they are in the data center,
like they are in a phone.

And that opens up the
opportunity for more more

vendors and more,
you know, complexity,

but but also
opportunities in software.

Yeah.

So that evolution of how
hardware and software is being

integrated in a new way brings
new opportunities to evolve the

relationships between
OEMs and tier ones.

How are you seeing that
relationship change?

Yeah.

So as we discussed previously,

it was more of a
straightforward relationship

right between tier
ones and OEMs.

You would buy a box,
you would integrate it,

you would test it, validate it, and
and and produce it, right, in scale.

In the future, I

there's going to be more,
more moving parts to manage.

Right. Right?

I find that the relationships
are now changing after

being kind of at a standstill.

Right?

So people didn't really
know how to navigate around,

recovering some, investments
that they had made.

So, you know, to develop an ECU
cost fifty plus million dollars.

Right? You have to
set up manufacturing facilities.

You have R&D nwork. You
have chips that you source.

There's all sorts of
work that the tier ones,

had to do to produce a product.

And then overnight,
OEMs changed their mind.

Right?

They wanted to, source,
hardware separately from the

software, and it took some time
for us to kind of work through,

how to decouple,
that that software.

But we see now,

there's a bunch of
announcements that that we've

seen recently where software is
now decoupled from from hardware.

Right?

So that, software as a
product is something that you

can purchase as an OEMs,

but then the complexity around
that is then finding how to

integrate all of that together,

and then test that in a
virtual type of environment,

and

yeah, try o find that
that right balance.

Right?

That's where I try to also position
Capgemini as a software expert.

So it's going to take some time
for OEMs to figure out how to

operate as a as a tech company.

Right?

And that's where engineering
services providers are

perfectly positioned.

Right?

Because we have the experience
of doing this already. Right?

We've been doing this for years.

So we're seeing, you know,

a lot of movement in
the industry where,

we're being approached —
engineering services as a

whole, right — are
being approached,

to kind of help bridge the
next, you know, whatever it is,

five years, ten years, to
do that work on behalf of

the OEMs while they kind
of not stand by and watch,

but they learn and then start to take
bits and pieces over to do themselves.

Right?

It's a journey.

Obviously, OEMs are interested in taking
a more active role in their software.

But it doesn't happen overnight.

And I think the learning of the
past couple of years has been

that it it is good
to be more involved.

It is good to be
more aware of that.

But, it can be a
stepwise process.

Yeah. For sure.

Yeah. Yeah. I I you mentioned
that this evolving, landscape.

I think we would be remiss
if we didn't observe that the

landscape's also evolving
incredibly fast in China as well.

Chinese OEMs are doing
incredibly innovative work.

The design cycle's faster.

They're coming out with
really incredible technology.

What are the some of the
opportunities to to learn from

China and some of the
challenges? What do you see?

So I think there's there's a
couple different things here.

So the one thing is, you know,
decisive leadership, right?

So,

fail fast approaches.

Right?

So,

launching a product
onto the market,

not knowing exactly what how
all of the pieces of software

are coming together,

but having fundamental
technologies like over-the-air

updates to where
you can quickly,

bring those pieces of software
up to speed for for production.

And also with with that
decisive leadership, right,

there there's there's no
opportunity for things to get

stuck in, building a
general consensus for,

and and and in the
middle management.

Right?

So it's more of a, you know,

here's the direction that
we're going to be moving in and

everybody executes as opposed
to finding more of a democratic

way as we currently
employ in the West.

Well, it's a couple
different parts of that.

Let's pull that apart.

One is my observation is
I think the Chinese OEMs are

looking really carefully
at the user, at the driver,

and saying, what do
the drivers expect

based on, for example,

what they're seeing with their
other digital device or phones

and other things like that?

And and it feels to
me — my perception,

I'd be curious if you agree
— is that they're delivering

technology that's much more
like other consumer devices

into vehicles in
a more aggressive,

more proactive way than other
is that your take as well?

Yeah. Absolutely.

I mean, you see that kind of
in everyday life in China as well.

So, there's a lot more convenience
as you go about your day.

It could be Alipay or,
you know, WeChat or, you know,

any similar type of technology.

So I think the the consumers
themselves are more demanding.

Right?

And so they kind of expect
a vehicle that they buy,

either develops with
new features over time,

and that's something
that's quite surprising,

especially for,
the the US market.

Right? We we also like
shiny things. Right?

And, and and Tesla
really brought

that, right, years ago.

And so I'm quite
shocked to see that,

we're still here all of
these years later with traditional,

and incumbent OEMs.

But, I think the yeah,

this this will change.

Your sense from the market is
that the rest of the world is

gonna evolve to have those
similar kinds of demands?

Absolutely. As a
driver, as a consumer.

Yeah. Absolutely. Yeah.

That's my that's my
personal take as well. Yeah.

I'm watching that
space closely. Yes.

So on that second point
you mentioned about,

the learning from China is
Chinese tend to have a more top

down approach versus Western
OEMs that might have a more

consensus based approach
to their development.

What are some of the learnings
that OEMs can take away from

the Chinese approach?

I think,

also not not moving so
finding the right speed.

Right? We we we have talked
a lot about the China speed.

What is China speed? Right?

It's a perfect balance
of hardware and software

solutions, right, that are meeting
what the the market demands.

And what I'm seeing with the
Chinese is that they're they're

not moving to
full SDV solutions immediately.

Right?

So a a lot of Western
OEMs are looking at,

centralized computing, HPCs,

Zonal controllers, while
the the Chinese have,

you know, they have,

fully capable domain
ECUs paired together with

centralized computing, but
there's only one computer.

Right?

So that's the way that
they can kind of balance,

cost while also
meeting market demands.

And I think it's going
to be a slow but gradual,

evolution for them, right,

to then eventually move to what
we see as zonal architectures

with

zone ECUs and, all of
this kind of stuff.

Yep.

So it's really a journey.

The the evolution of vehicle,

vehicle EE architectures
is not a single direction.

It's not everyone is going
exactly a certain way.

There's a lot of exploration.
There's a lot of differences.

And, but what you're saying is
there there's they're really being

willing to try new
things and evolve.

Yep.

And I think one one
thing that's really,

fundamental for this is that
they have they're they've

already achieved,

different relationships with
their tier one suppliers.

So when they saw when
they source a box,

the software is already
white boxed to them.

Right?

So they are able to really
understand what's in there.

They're able to develop
upon that software,

and that is also allowing
them to then move software.

Right?

As they want to progress
their EE architecture,

introduce additional
ECUs, move things around,

that's going to allow them to,
to achieve that much easier,

right, without any pain points..

Well, so if we pull back
to the worldwide OEMs,

we we've talked in a couple
different ways about,

the evolving architecture.

What about the,
organizations in in OEMs?

What's your advice based on
your you're talking to many

different companies.

What's your advice for how OEMs
could kick off an SDV project?

Yeah.

So I think, we've
actually seen a lot of,

great organizational changes
already, big, on the back of,

electrical vehicle development.

Those models were
were great, right?

They didn't work for for other
reasons more from the market

perspective, right?

But I think a similar approach
where you can kind of have a,

you know, small tugboat
size, organization, right,

that with a bunch
of heavy hitters,

people that really understand
all the technologies,

through and through, able
to really develop that

first SDV, right,

and then get that learning into
the company and then use that

to then drive organizational
changes further.

Right?

So there are certain parts that
don't need to change. Right?

Mechanical parts are always
gonna stay mechanical,

but then from the the software
reorganization, right, I think

small targeted expert
teams are really the key.

So it's really a
cultural shift, really,

of changing away from,

a very verticalized silos,

stay out of my silo
approach to really

thinking about design
in a in a different way.

We've seen that in other
industries, in data centers,

in in in computing,
other parts of computing.

We need to bring that kind of
different thinking into automotive.

Yep.

So previously, you know,

you would have people that
were responsible for one little

screw, and you, you know,

they weren't weren't able to
even work outside of that,

that that box.

Right?

And I think really
cross -systems thinking,

almost turning

the responsibility of teams
ninety degrees, right,

instead of a vertical approach.

It's more of a
horizontal. Right?

Assembling teams that
have mixed skills,

so it could be IT skills,
it could be people,

with finance backgrounds, right,

payment systems and all this
kind of stuff paired together

with engineers that understand
how a feature works and data

flows work, across
an entire system.

Yeah.

I think it's a different
mindset and more of a holistic

systems approach where, yes,

there's division of subsystems
because even in traditional

computing and phones
and data centers,

there's obviously lots
of division of labor.

There's lots of walls that
prevent apps and compute from

messing with other compute.

But there's also the
realization that those other

things are there and thinking
about it in a holistic way.

That's, I think, the difference.

So as these vehicles evolve,

SDVs provide new ways for
OEMs to differentiate.

So, tell me about your
opinion of how they should be

differentiated and where in
the compute stack should the

differentiation happen?

Yeah. So, yeah, great, topic.

I love this topic.

So we've been having a lot of
discussions on this, right,

for many years, right?

You'd have, OEMs go to
conferences and say, you know,

we need to collaborate,

on things like middlewares,
OEMs, different commodity, items.

Right?

And then everybody would go
back off and invest hundreds of

millions of dollars
in their own solution.

Right?

And I think that it was kind
of this similar to, AD. Right?

A lot of people wanted
to be first to market,

without recognizing all
of the challenges, right,

that we're going
Autonomous driving.

Yeah.

Exactly. Yeah.

I mean, both
autonomous, but also,

middleware development programs.

We we we saw a lot
of, issues there.

I think that over time,

the the all the entire software

stack kind of, middleware
downwards, right,

can be and
should be commoditized, right?

So anything that the
customer doesn't see,

they don't really care about.

They don't know that it's there.

And I think that if companies
come together and co invest,

there's a lot of cost
savings to be had, right,

which can then maximize
investments in other areas, right?

Differentiating products in
digital services offerings and

unique experiences across
different types of, products there.

One thing that I'm happy to
to report is that you see this

already today in the
commercial vehicles, industry.

Right?

I find that they kind of
they're taking all of the right

learnings from passenger
car and applying that

in their own, industry,

partially driven because
they have smaller and tighter

budgets to work with.

Right.

But really, it's the right way,

regardless of what
your budgets are.

Right?

I think it's really, the
collaborative approach,

open source software,
and really commoditizing,

the middleware downward, layers.

Yeah.

So this the idea from a
passenger perspective,

the ability to differentiate a
particular vehicle model or a

vehicle family, whatever,

to provide an experience
that's unique to your brand.

I mean, you see, you know,

many different brands who
have a very different user

experience, a different UI.

Even my wife and I have cars
that are very different,

and I won't name the names,

but very different
user experiences.

And she's very happy with hers,
and I'm very happy with mine.

But then you mentioned
commercial as well.

Commercial, obviously,

they are much more
diverse fleets.

You can have fleets
that are carrying cargo,

that are carrying passenger,
carrying cold chain.

They have to keep things cold.

And there are there are very
specific vertical market

differentiating needs there
to serve those unique needs as

well as large large vehicles
and smaller vehicles.

So it's really I think it's a
valuable point you make that

there's there's lower levels
that are non differentiating,

provide those in a common way,

but then still retain valuable
differentiation for the OEMs and tier ones?

You know, with SDV,
a lot of times,

the perception is SDVs
are for high end vehicles.

And indeed, some of the initial
SDVs have tended to be higher end

vehicles where the
first innovation is happening.

But then naturally, I think
it it can move down into a

wider range of vehicles as well.

I'd love to hear your
perspective on the

applicability of SDVs
across the price points.

Yeah.

I mean, I think, you know,

we can touch on this in
in two different ways.

Right?

So, specifically in the
commercial vehicle space,

right, you have, light, medium,
and heavy duty vehicles.

Also in the passenger car space,

you have entry level vehicles
all the way up to the full

singing, dancing,
large vehicles.

Right?

Today, that's a difference
between maybe, I don't know,

six and a hundred ECUs.

Right?

So, if you look at that
from the SDV perspective,

I think it's it's important to
find a minimum set of hardware

that's needed to,

deliver a certain feature
experience and and feature content.

But I think that, you know,

if you compare that
approach, right,

with partnering with
people, commoditizing architectures,

increasing piece
purchases, right,

so that that lowers piece
cost, you're able to then,

find ways of delivering
more expensive

single pieces of hardware
into cheaper vehicles.

Right?

So another idea,

that that I've been discussing
with a lot of people is to

have, you know, one single
piece of hardware with,

firmware limited
processing power.

Right?

Instead of having different
across the chip suite,

instead of having a physically
different processor in there,

you could have the same processor
that's firmware limited,

and then try to
find a way of paying

for processing power as and
when you need that right as

feature content grows as the
market demands additional

features you can
then find agreements

with either silicon
silicon vendors either

via the tier one or
directly with them,

to then pay pay per per use.

Yeah.

A couple of different points,

of what you made is
really interesting.

One is there there's a
a singular focus, I think,

on BOM cost, when
the reality is also

validation cost and and
development cost is a huge,

huge investment, a massive
investment for new vehicle models.

So if one if a company was
able to take a common platform,

that can scale across high
end vehicles, medium vehicles,

low end vehicles, then some
of those development costs,

which are significant,
can be reused.

Validation costs — validation is also
a huge component — can be reused.

So that's the first point.

And then the second
point you made,

which is really
interesting, is today,

a lot of times
chips are, you know,

fuse-burned to be a specific model
or SKU as we call it in the industry.

But there's an appetite and
there's an interest in having

scalable systems that can
add capabilities over time,

that can have
subscriptions over time.

And so I think the industry
is still learning how to get

through that approach of
having more scalable OEMs,

knowing that it might not be
using the full capabilities at

the beginning, but that
there's headroom to grow.

And it may benefit everyone,

including the silicon
suppliers who may have fewer,

SKUs to sell. Yeah.

Which potentially
could be beneficial,

and for the OEMs,

fewer boxes they have on
the shelf that are more scalable

across their product line.

I think the challenge for for
them is going to be estimating,

you know, over the
lifetime of a vehicle,

what is the the maximum
amount of processing processing power

that, can be
requested by the OEM.

Right.

So forecasting, you know,

ten years in the future
is a challenging thing.

It's challenging. Yeah.

We just recently released the
results of the 2025 SDV survey

we carried out with
Wards Intelligence,

actually one of the, key
partners of Wards Auto.

So, it was interesting to see how
much headroom OEMs were willing

to put in their vehicles.

And it definitely varied.

Higher end vehicles were
willing to put in more headroom.

But even in lower end vehicles,

we saw an appetite to
include headroom in vehicles

recognizing that there's
a need to scale over time.

So it's an
interesting challenge,

and I think it's gonna be an interesting
topic to watch for in the future.

As people are paying for
more pay-as-you-go and more

subscription services,

I think there'll be more of
an appetite to put in headroom

into these vehicles.

Well, look, Suat,

it's been so wonderful to
have you on the podcast.

You're you're such
a thoughtful leader.

You work with companies
of many different types,

so it was exciting
to have you on.

I've been working with
Capgemini for years,

and so thank you for
being with us today.

Thanks again.

If you like what you're
seeing on this episode,

please like and subscribe to
see more episodes like it.

More from, AutoTech here in AutoTech in
Detroit as well as from our home studio.

We'll see you in
another episode again very soon.