The Garage by Sonatus

Today in The Garage,
we're recording live at AutoTech

Detroit.

My guest is Joe Capuano,

who's head of technology
for Bosch Mobility Americas.

Joe has an incredible thirty
seven year career in the

automotive industry and
thirty years at Bosch.

In today's conversation,

Joe shares his perspective on
where we are on the journey to SDV,

the importance of system level
thinking across a range of

these vehicle
design price points,

and how AI can help us manage
the growing amount of data that

vehicles and SDVs will create.

I think you're really
gonna enjoy today's conversation.

Let's go.

Welcome to The Garage.

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

We're recording live at
AutoTech Detroit in Novi,

Michigan.

That's the background noise you may
be able to hear during the show.

Our guest today is Joe Capuano.

Joe is Head of Technology
for Bosch Mobility Americas.

Joe, welcome to The Garage.

John, thanks for having me.

So, we always like to have our
guests introduce themselves and

tell us about your background.

Tell us about you.

Sure.

I'm responsible,
as you mentioned,

Head of Technology for the
mobility sector in Bosch or

Bosch Americas.

I also have a second role,

which is I've been doing
for four and a half years,

which is, Head of Bosch
engineering, North America.

So responsible for all
the Bosch products,

but selling mainly to smaller
customers or in smaller volumes

and some other niche parts
of the mobility industry.

My background, I'm
a computer engineer,

have worked in the
automotive industry,

thirty seven years now.

Thirty years at Bosch.

Congratulations. Thank you.

Doing, basically working in
engineering of all different

automotive systems that you
can imagine, powertrain,

chassis, convenience systems,

very heavy, effort
and emphasis on

software, electronics
and software,

digital electronics and software
that goes along with it.

And have,

launched several product
products into production with

our North American customers.

I've also spent five years
in Europe at our Germany

facilities, at Bosch,
learning the culture,

working together with European
customers and European,

my European counterparts.

All in all, been a great run.

That's fantastic.

You know, one of our Sonatus
team members, you may know,

worked with you in the past.

And when we saw you were on the,

roster for this
event, he said, oh,

you got to have
Joe on the podcast.

So we're thrilled to have you.

You got to also start us off
with a fun fact about you.

Yeah.

So, I'm an avid,

two-wheeler guy.

I love motorcycles, bikes,
mountain bikes, especially.

And over the course of
the last fifteen years,

I have gone coast to coast
from Newfoundland all the way to the

edge of Canada and all
the way to Prudhoe Bay,

Alaska by motorcycle Wow.

Without staying in hotels.

So with the backpack and a tent
and kind of roughing it a bit.

So That's incredible.

A very big adventure and it was
kind of a dream of mine to do that.

That's phenomenal.

You must have seen the
Long Way Round series.

Oh, yes.

Of course.

That's an amazing series and
so I feel like that's the

adventure you lived.

So hats off to you.

I always I always like to say
something to match our guests.

I have no experience
motorcycling

and I've I definitely have not spent
as much time in Canada as you have.

I feel bad about that.
I should do that.

But I used to be a
cyclist and a triathlete.

So I did that for some
years and that was,

that was a great experience.

So tell us about, Bosch
in general for our guests.

I mean, it's a famous company.

So but just a little context
about Bosch in general and then

also what do you work on?

So Bosch is the largest
automotive supplier in the

world, a hundred
twenty years in the US,

nearly sixty billion in
revenue in automotive in 2024.

So, a full range of portfolio

products represented in every
automotive region in the world.

Anywhere vehicles are made
or sold, you find Bosch,

you find Bosch service,
you find Bosch parts,

aftermarket parts, a very
broad portfolio of components.

We're supporting nearly every
customer in the world in one

form or another through
our products and services.

Manufacturing is a

key core competency at Bosch,

but software is also now a
core competency at Bosch and

something that we're taking
very seriously as we move into

this SDV realm in the future.

Great.

And what sorts of products
do you focus on yourself?

So I personally have worked
the longest in powertrain at Bosch.

I was, responsible for our powertrain
controls business unit for over

ten years in the US,

and that's one of our prime
product pieces out of the

product portfolio.

We provide full fuel systems,

but the controls of the
controller and the software are

also very critical to the
functionality of the systems.

And so that was a
big area I was in.

I was also
in safety electronics, airbag,

occupant sensing.

And in my role at
Bosch Engineering,

that I've been in the last
four and a half years,

we support all the products
of the Bosch portfolio.

So it's a very unique,

role that I have and a very,
very unique job because,

essentially, any hardware
that Bosch makes,

I can apply to smaller
customer volumes.

So we take off the
shelf hardware.

We are able to customize the application,
the calibration, the software,

and then provide the benefits
of high volume electronics to

smaller applications.

But we do that across the
full product portfolio.

So I see all of
the Bosch products.

That's great.

That puts you in a really,

great position to see
the whole landscape.

That's really fantastic.

So I wonder, we often talk about SDV and
software in this podcast, and in general,

it's important to Sonatus and
important to the industry.

We were chatting about SDV.

I'd love to get your
take on, you know,

where we are in the SDV journey,

and how are things going?

To me, SDV is more like a continuum
of actions that have started with

the with the advent of
ECUs in the vehicle.

So when I first started
working in automotive,

I was working on engine control modules
that could not be flashed at all.

So they only had a one time
program programmable memory,

and they worked very much
in a localized domain.

So there were sensors, but
there was no high speed bus,

activity necessary because
the system was standalone.

It did what it had
to do. It met regulatory requirements.

Right.

But it was a single
piece in the vehicle,

and probably the vehicle just
had a couple of things like that.

Sure. But then additional
functions came along like ABS.

And all of a sudden,
once we had ABS,

we could control wheels.

The two systems,
powertrain and ABS system,

needed to talk to each other.

So by expanding
the functionality,

we wanted to add traction
control into our ABS system.

We had to be able to
manage engine torque.

And that initially was not done
via some complicated CAN bus or

something like that.

That was done with
pulse width modulation.

Really?

So there was a torque requested,

torque received signal, and
that's how the communication went.

But you can imagine already
people were thinking,

couldn't we somehow combine these
two pieces that we have together?

But the technology
just wasn't there.

Right.

The microcontrollers,
the memory,

it was extremely expensive back
then to add a lot of memory

into, a lot of memory and
computation power in the systems.

And to a certain extent,

the take rates weren't
always supporting going to a

completely new architecture.

But I would say the writing was
already on the wall early on

that someday we're gonna get to
the point where we bring things

together because it makes
sense from a take rate, from,

a complexity, from a
pricing standpoint.

And we're kind of
at those days now.

As we have over a hundred
ECUs in many vehicles,

it's just impossible to
continue at that pace,

with such a distributed network.

You have to look at centralized
opportunities. Right.

And really the big motivation
for this is autonomous.

Because once you
have autonomous,

nearly everything in the
vehicle has to be able to be

controlled by a
central computer.

And to do it all via buses,

even Ethernet and very high speed
buses just becomes too complex.

Oh, yeah.

We were talking about,

the optimism versus
the reality, of this.

And indeed, autonomous driving is
an area I'd worked on before and you

mentioned you've been worked on.

It's still something
that's very much evolving.

There's fantastic progress.

You know, Waymo and companies
are doing incredible work.

Obviously, L2+ / L3 systems are beginning
to roll out from the leading edge vendors.

But still a lot of
it's in front of us.

Maybe we could talk about the
optimism versus reality in SDV

progress and various aspects.

Yeah.

I think on the autonomous
side, it's coming.

But none of us can put our
finger exactly on the timeline

to say when that is.

But SDV needs to be there
to support even the advanced

driver assistance functions
and safety functions.

For example, automatic
emergency braking

is a function which
is becoming mandated.

NCAP required as well.

Yeah.

And so you have,

you have to combine the data
from several systems to be able

to do the things that you need to
do to support that functionality.

So it's I see the roadmap
as an evolutionary.

Of course, there's a revolution would
be level five autonomous comes and,

it came it comes very quickly and we
have to make a complete paradigm shift.

But really more what's
happening is that it's a

gradual increase with certain
functions like AEB becoming

mandated, becoming mainstream
across the application.

Yep.

We're taking a step toward
looking at opportunities to

combine functions
together in that case.

And we can't ignore the
connectivity issue either.

The vehicles are just becoming
more connected because of,

additional convenience
functions and the things that

we talked about in
the panel today with,

you know, the need to
have data available to,

develop new vehicles.

There's marketing
and, obviously,

there's other aspects with respect
to what's done with the data.

So connectivity and a AV together
are really pushing driving,

the topic as well as just the number
of ECUs and complexity of the vehicles.

Right.

So imagine a skeptical
listener says, oh, okay.

AV is gonna be in the future and I
don't need to worry about that today.

The reality is, as
you said a minute ago,

that there's still a need to
solve these problems today Yes.

For passenger vehicles. So
today and probably trucks too.

I mean, we often
talk only about cars,

but the benefits to trucking
is equally important.

So if we talk about
non AV applications for a minute,

let's explore some of the
benefits you're seeing with,

you mentioned
hardware consolidation

that can be a cost savings
and a simplification.

You mentioned wiring.

What are some of the areas you
think are the most promising

short of AV for SDV?

Yeah.

I think the digital cockpit
is obviously, a component of,

the SDV, the need for SDV.

We have a lot of, decentralized

control units, in the vehicle.

And I think there's
opportunities, like,

in motion control to bring, functionality
from power train, from braking,

and from other domain steering
and other domains together,

into motion control.

There's vehicle
integration platform,

which is a Bosch product where
we're bringing more vehicle

functions together as well.

These are operating,

at the next level up in
terms of the architecture.

They need to share data,

and communicate to the outside
world very differently than

what we see from a
decentralized component,

which is, an actuator or
close to the actuator.

Right. In all of this, you
have to consider the system.

And I'm a big believer in
doing good systems engineering.

You have to look what
you want to have,

what is inside your system and
what is outside your system.

The simple example I always
give is an accelerator pedal.

If you say the accelerator
pedal is my system and you draw

the circle around it,

the input is someone's foot
pressing applying force to the

pedal, and the output are
voltages on the potentiometers.

But to the consumer, the system
is I press my foot on the

pedal and the vehicle drives.

If I draw the box like that,

I have to include the complete
propulsion system inside the system.

That changes the system
challenge I have,

and it changes the architecture that
I would approach the system with.

Exactly.

I mean, that it changes
the scope of thinking.

And I you mentioned the so many
hundred ECUs and something like that.

And on the one hand,
you could argue, oh,

that's a really smart
division of labor.

See if we can do it later.

But it's no one would design it
that way if you started with a

clean sheet of paper
with today's technology.

So as you look at and, you know,

you're talking to so many OEMs and you
have such a great view of the industry.

There are many different
choices about how future

vehicles can be architected.

Just yesterday, you
may not know this.

Just yesterday, we released the
results in collaboration with the Wards

Intelligence of the
2025 SDV survey.

Looking at, you know,

key industry leaders
around six hundred industry

leaders talking about their view
of where vehicles are evolving.

And what you saw was
a range of, ideas.

There's no single architecture. There's
no single view of that.

So I'd love to take get a
sense of your survey of one.

What's your sense of where do
you think the industry is gonna

go as the predominant
direction ranging from today,

largely totally decentralized,

to one big computer in the middle
and no other computers there.

Probably the answer is
somewhere in the middle.

Where's your take on that?

The answer for
unfortunately, is it depends.

So, if you take customers,

as you mentioned,

that are designing really with
a smaller legacy footprint,

from a clean sheet of paper,

they can be very aggressive
in terms of how they optimize

their architecture.

Their system is smaller in
terms of requirements and

legacy needs than someone might
have who has been producing

vehicles for a
hundred years Right.

Or more.

So the I think we're gonna see,
a lot of different solutions.

It's gonna be a stepwise approach
for the larger customers.

They can't abandon some of
the components and some of the

legacy that they
have from the past.

They can't make a jump to a
brand new architecture and

leave everything behind where
they still have a lot of

profitable product potential,

but they need to make
the step in parallel.

And I talked about
it in the panel.

It's not some of these non
regulatory features are not

gonna have hundred
percent take rates.

So architecting a system
which has a lot of expensive

components in it in terms of
compute power and distributed,

distributed ECUs that
has to cover also a lower

end vehicle that will
add cost to the vehicle.

And I see an overriding
theme I see in my personal

life as well as at Bosch is the
affordability of transportation.

And we have to keep
in mind that, yes,

some of these systems,

are very there there's a lot
of creature comforts available,

but they're not required
to have mobility.

And our systems have to be
flexible enough to be able to

deal with the
lower end vehicles,

but still support the higher
end vehicles without having two

totally separate
electrical architectures,

which is very expensive
for our customers to carry.

Yeah. Yeah.

That last that last point
is a really important one.

We've had other guests this
week talking about exactly that

that one of the potential
promises of SDV is you can have

a common architecture and yet
intercepts at different price points.

So the validation cost,
which can be significant,

the design cost, engineering
cost can be reused.

But recognizing that some will
be more outfitted, as you said,

take rate than others.

Do you see that as the kind of
the direction of travel where

there's a a common
framework and a common

architecture but
different price points?

Yeah.

This is not the way
it's done today.

Yeah. This gets into, a
question of scalability.

Right.

Right?

So,

if in the past, the way
that was done was, you know,

you had microcontrollers,
generally speaking,

in our in in our,

separated ECUs or
distributed ECUs.

And, you had a
microcontroller family,

and you could scale
the microcontroller and the

software to fit the
application within a family.

We need to see that same
level of scalability at the

microprocessor level,

what you're operating at the
vehicle control unit level.

And then in this,

more common, with
a lot of common

inputs, a lot of non
distributed functionality

like, yeah, some of the
functionality we spoke about before.

We need to have the scalability
to enable us to provide lower

end solutions within the
same architectural framework.

And the second aspect, I talked
about it as well, standardization.

Yep.

So, the interfaces need to be
standardized enough that we can

use those interfaces both for
lower end applications as well

as higher end applications
without causing a tremendous

amount of additional work for
the engineers of validation and

integration and the things that really
drive development timing and cost.

Those two things need to come
together. I think they will.

So I think, and then we're
seeing signs of it already.

There won't be a big bang SDV,

at least not that won't be the

mainstream solution, I believe.

It'll be an
architectural evolution.

Some elements will
be brought in.

Some scaling, components
will be brought in as well.

And the next generation will
bring more in and bring more

flexibility and more
standardization.

And finally, we'll end up eventually
in an architecture, which is,

maximizes the amount of integrated
components in the system.

So a couple...you mentioned
such interesting points there.

One is that the importance of
having scalable performance points.

So you have a processor or,
you know, vehicle processor,

but that has higher performance
and lower performance

offerings, but are
software compatible.

Yes.

So that's one of
the points you made.

I'm wondering where do you come
down on the kind of the zonal

architecture consolidation?

Because today, you have
these end sensors, you know,

distributed parts of vehicle,
whether it's ABS or, you know,

ultrasonic or, you
know, whatever they are,

around the vehicle.

How much of that do you
expect is going to get to consolidate

into zones, so that you
still have software,

but those microcontrollers
or, you know,

combination microcontrollers,

microprocessors are in the
quadrants of the vehicle.

Do you see that as a
direction of travel?

Zonal architectures are a big
discussion point right now,

and, several customers are
moving in that direction.

exactly which sensors and
which functions are integrated

varies, I would
say, by customer.

And this is a
challenger to suppliers.

We need to still provide
discrete solutions,

but we also have to provide
integrated solutions.

And the software can't
be totally rewritten between

discrete and integrated.

So we need we need
a kind of API.

Kind of a graceful transition.

We need a graceful transition.

But as a supplier,

I'm pretty certain we're going
to be required for a relatively

long period of time
to provide both.

Yeah.

We can do that in
a smart way Mhmm.

By defining the
APIs very you know,

by really getting
into the middleware,

that we talk about within our
zonal architectures and making

sure that the software is
written with standards so that

we can utilize the software
from the discrete side in the

zonal side as well with
a minimal amount of,

integration effort and a
minimal amount of adaptation.

So every customer is gonna,
draw their, you know,

circles differently in terms of what
gets integrated and what doesn't.

It's also highly dependent
on bandwidth of the data.

And if you want to
bring, you know,

the data from a radar
to, the a central ECU,

you're gonna have to do some
preprocessing at the at the source.

Right.

Which is basically
another discrete ECU.

Right.

It's just the bandwidth
of the data OEMs,

in some sensor technologies,

becomes prohibitive to be
able to have just one ECU with

sensors around the vehicle.

Right.

You mentioned some different
networking technologies,

you know, that they you
know, obviously, today,

proliferation of CAN and, many
other solutions like that.

How much do you see the take
rate of automotive Ethernet as

the next generation backbone?

Because it's such it simplifies
the wiring harnesses so much.

Yes. But it's a big change.

What's your take on
the timeline of that?

I mean, it's an integral
part of the SDV architecture.

So, I think you'll
see that coming,

in many different applications
in the next few years.

That's great.

You also mentioned
standardization, both, you know,

standards bodies and also
standardization interfaces.

Are there particular
standards or standardization,

bodies that you think are
particularly important and relevant?

Yeah.

So we spoke about COVESA.

That's obviously, an
important aspect in SDV.

I mean, AUTOSAR or
AUTOSAR Adaptive,

are continuing to
develop standards,

which are critical
for SDV adoption.

There's also a lot of
standards at the, cloud level,

that are being, brought in.

We had discussions about those
during the panel discussion as well.

These are all important aspects to be
able to get to a real SDV

architecture, and they're
gonna continue as well.

One of the things that came
out of the SDV survey we talked

about is, a growing
appetite for,

CI/CD and virtual prototyping.

That's one kind of, if you will,

corollary benefit that
comes along with SDV,

that can also, in a in a
sort of virtuous cycle,

help to accelerate the adoption
of SDV and bring other benefits

like cost reduction and so on.

I know that that's a key,

something you care
about as well.

Can you talk about your take on
virtual prototyping digital twins?

Sure.

Really, the all these
issues are tied together.

The more data that we get from
the vehicles also from the field,

the more that can feed back
into our development process

and improve our future products.

So this is, I would say, SDV
is an enabler toward more data.

Really, in the end, we're
getting so much data now.

It also becomes an issue.
What do we do with the data?

And do we have the right competencies
and people to work on that

data?

And we're

working very hard at Bosch.

We have an AI training
institute, within the company,

which is focusing on software
for all levels, you know,

executive all the
way to our new hires.

We're working on, you know,

kind of volunteer

types of working groups within
the company where people sit

down and work with
different tools,

Python and other things to be
able to work and look at what

the environments look like.

So we have a very experienced
group of automotive

professionals at Bosch,

but not all of them have been
trained in AI and AI tool sets

and methodologies.

And so we need we need
that domain knowledge.

It's absolutely critical to
Bosch's success and I believe

the market success that
automakers and tier suppliers

maintain their automotive
domain know how,

but they need to bridge into
this tech world where we talk

about AI, AI development tools.

And, really,

AI has the potential to not
just in the prototyping part,

but the entire
development cycle.

There are tools out there now
to facilitate that and make

better use of model
based development.

The days of needing to sit down
and code anywhere in the world

are slowly coming to an end,

and we have to really operate
at that next level of AI

integration where the tool sets
are developed and our engineers

are trained to take advantage
of everything which is

available to them.

Yeah. AI is such
an important topic.

First on the narrow point you
made and then I'll speak more

broadly about AI,

you mentioned about the importance
of using AI for data mining.

That's not the exact words you
said, but that was the idea.

It's interesting.

I'm just I'm smiling because
right behind that wall,

we're showing a demonstration
of exactly that,

how we've come up with a
solution as well to seek that

same idea of helping,

OEM engineers to be able to
do more detailed queries with

natural language more easily.

Yes.

Instead of having to program at
the low level and know exactly

which weeds level signal.

That's crazy. Why don't we use
AI just to accelerate that?

If I pull back
more broadly on AI,

one of the other conclusions of
the survey I mentioned earlier

was, looking at we asked
our survey participants,

what are the most promising
applications of AI in vehicles?

And it's so interesting
the results.

Four of the five of
their top answers,

were things that are
largely not deployed today.

Number one is
vehicle diagnostics.

Using AI to find
and fix problems.

And number two, I believe was
vehicle tuning and optimization.

Things like using AI to
do ongoing performance

optimization, efficiency optimization,
and so on.

Vehicle personalization
was one of them.

And number four was,
okay, IVI and navigation.

You'd expect that. Mhmm.

But so fully four
of the five top,

applications were
not deployed today.

So huge opportunity

to deploy into the vehicle.

And then you made a broader
point about AI as a tool.

You know, AI is of
course, it's a, you know,

big buzzword today.

But the fact is it's such
a powerful capability

for so many things,
improving productivity,

improving design cycle,
improving effectiveness.

It's exciting times to
watch how it's being used,

across the space.

Yeah. Absolutely.

You mentioned,
also that you have,

you do with large customers.

You do with small customers.

I wonder if you can share some
of the learnings you've had

about some of the things you
had to do differently with

large and small customers.

Yeah.

So most of my Bosch history is

working with large customers.

So, in power train
combustion, for example,

there's a lot of large
customers in the US that that

we work very closely together
with, on software and,

some of the topics
that we talk about,

but it was more in a
traditional architecture.

Yep.

When it comes to
newer customers,

there's obviously a lot of
flavors of new customers.

Most are very eager to
make rapid advances in

development to get their first vehicles
out to market as fast as possible.

And so they're not

always so interested in finding the
optimal electrical architecture.

They want off the shelf product.

They want it adapted
for their needs,

and they want it as
fast as possible,

meeting all the
requirements that they have.

But there are also customers,

that we work with that also are
working more on a blank sheet

of paper side and say, okay.

We have a lot of flexibility here
in latitude and what we can do.

We have kind of a our guiding light
is where this is the direction.

This is the DNA we want
our vehicle to have.

But let's work together on what
you can bring to the table in

terms of advanced architectures to have
us really be able to do

the types of things that we wanna do
with the vehicle user experience wise.

So very fast, you know,
flash over the air cycles,

continuous improvement.

You know, this
customer expectation,

the vehicle on the first day is
at its worst rather than its best.

Right? So we want
improvements over time.

We can build those things
in for those customers.

And, obviously, those are big issues
that we deal with as well with our

larger customers.

But many customers, as I said, are
just saying, get me into, you know,

production as fast as I can so
that we can have product on the

market and we can begin
a cash flow stream.

So we're dealing
with both extremes,

in terms of what we
deal with in the market.

We've been talking a lot
about, or implied, passenger vehicles,

but I assume a lot of these
same trends apply to commercial

vehicles as well.

You get exposed to
commercial vehicles?

Yes.

And what are some of the issues
you see there that may differ

or be common?

Yeah.

I mean, commercial vehicles
have different set of regulatory

requirements, so there's
differences there.

Obviously, volumes are lower.

We deal also with
off-highway vehicles and, the

SDV,

over the air topics are
really prevalent there.

So smart farming, for example.

And, you know, cloud connection
is critical to getting,

you know, data about
the fields and,

you know, lining your tractors
up and rowing and mowing in the

proper, spaces and, you know,

treating bugs and
things only where it

really needs to minimizing
the use of pesticides.

They're actually quite advanced
on the off-highway side.

I would say even more so
than what we deal with on the

passenger car side.

So these SDV elements,

central computing and
connectivity systems are

prevalent in
off-highway especially.

And we also deal at Bosch
Engineering with rail.

So we really deal on the,
some of the nontraditional,

mobility side with,
safety systems.

So, we have our tram
collision warning system,

forward warning system.

We have activities in,
other spaces as well.

I mentioned off highway.

We're selling components, radar,

and other components
in off highway.

So we're getting exposed to
how the different customers and

different segments are
using our components and,

how they wanna architect
their system so that they have

maximum flexibility and
functionality that they request.

Alright. I'd love to pull back.

This is such an interesting point
you made about, off highway,

and I know there's a lot of
incredible innovation in off

highway sort of
farming and things.

You mentioned how they're using
OTA in a very advanced way.

One of the other data points
in the survey that was so

interesting was, while a lot of,

companies said that OTA is now
prevalent in their vehicle,

It's also clear that that
they're only using it today

for, I would say, largely bug
fixes and in a smaller way.

But what the survey said is
in the coming couple of years,

the adoption of using software
to add new capabilities in the

vehicles was expected to
accelerate much more rapidly.

The general trend from the
survey also was just that SDV

is moving away from it being a
buzzword to becoming more real.

And I think that one state
data point was a really key

statistic that said that's the tipping
point when you add new capabilities.

So as we talk to people,

I think sometimes
there's a reticence

to think, oh, well,

how can I add new
capabilities after shipments?

And I think that's
a huge fallacy.

I'm wondering what your
take is on that aspect.

Yeah.

It's a great question because,

the expectation that
we're kind of setting with

customers now is what I said before
is there's improvements over time.

But there there's a high
dependence still on the hardware.

Right.

To support that.

And generally speaking,

we don't build more into the
hardware than we need, right?

So the hardware is optimized
from a cost perspective and a

reliability perspective to do what
it needs to do in the application.

But if we're going to expect
the software to do more,

we may have to provide
additional functionality in the

hardware too to support it.

I'm talking not as much
about digital cockpit.

There it's clear with apps
and things you can bring new

functionality in
in different ways.

But if we wanna bring it to
mainstream, you know, portion,

you know, our brake controls
and things like that,

we have to make sure that we're
ready to do that with respect

to the hardware as well.

And that's I don't see
as much of that yet,

but I could see that coming.

But it's, it's we're kinda separated
with respect to those two domains.

I see that. Absolutely.

And we've had several guests on
the podcast talking about this

topic and how it's challenging
today with today's metrics of

success to build in headroom
because it bakes in cost

effectively with
uncertain revenue.

So from an accounting
perspective,

it's red ink from day one.

But one of the interesting
aspects in the survey we asked

was about, how much headroom
people can build in,

and there were various
answers across different price points,

low, medium, and high.

But I observe that as one
consolidates from many ECUs,

in which case to
build in headroom,

every ECU would have
to have headroom.

Two more consolidated ECUs,
fewer consolidated ECUs.

You have the potential
opportunity to share headroom.

So

in an aggregate, you have to bake
in less costs to keep headroom.

Do you see that as a viable
direction in one of the upsides

of consolidation?

Yeah. Absolutely.

I think that I mean,

the microprocessors that we're
building into vehicles today

are really fascinating,

compute platforms that really have
a tremendous amount of capacity.

I'm sure that our engineers will figure
out how to use it all eventually.

But, you know, from where I see today,
the differences between, you know,

the microcontrollers
of yesterday

and the microprocessors today,

there's really no comparison
in terms of what's available.

But it's more than just
compute power. Right?

So if you really wanna add
different functionality into

your system, you may need
additional actuators.

You may need additional
inputs and outputs sensors.

So it's not just a flash
over-the-air activity. Right?

It could be that you're adding
hardware to the vehicle as well.

So, I don't think compute is the

limitation, at least today.

I think it's more what do you
wanna do and what additional

function what additional
sensors and actuators would you

need in the vehicle to
fulfill those, requirements.

Again, setting aside digital
cockpit with apps and,

entertainment and
things like that,

which can be brought in
purely with software.

Joe, this has been an
incredible conversation.

You're such a huge resource
of having a great perspective

across the entire
system as you said.

I've really enjoyed the
conversation and we're really

grateful for you
visiting with us today.

Thanks so much, John,
for having me on.

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

please like and subscribe
to see more from us,

either from our episodes at AutoTech
Detroit or from our home studio.

We look forward to seeing you
again in another episode of The

Garage very soon.