April 4, 2025 • 28 mins
Karl Freund of Cambrian-AI Research, Jim McGregor of Tirias Research, and Dave Altavilla of HotTech Vision and Analysis joined Leonard Lee of neXt Curve at the Virgin Hotel in Las Vegas to recap @Intel's Vision 2025 event.
This episode covers:
➡️ Key impressions of the Intel Vision 2025 event.
➡️ The "New Intel" strategy of Lip-Bu Tan, Intel's new CEO.
➡️ Intel's AI strategy from a data center perspective.
➡️ 18A upon which Intel's hopes and future leadership hang.
➡️ Intel's less than compelling AI strategy and vision.
➡️ The diversity of AI and the heterogeneous present and future of AI compute.
➡️ The incremental Intel vibe of Vision 2025.
➡️ The fate of Intel Foundry - Killing the rumor.
➡️ The state of Intel's DCAI (Data Center & AI) business.
➡️ Intel Vision 2025 in a word.
Hit Leonard, Jim, Dave, and Karl up on LinkedIn and take part in their industry and tech insights.
Check out Karl and his research at Cambrian AI Research LLC at www.cambrian-ai.com.
Check out Jim and his research at Tirias Research at www.tiriasresearch.com.
Check out Dave Altavilla and his research at www.hottech.com.
Please subscribe to our podcast which will be featured on the neXt Curve YouTube Channel. Check out the audio version on BuzzSprout or find us on your favorite Podcast platform.
Also, subscribe to the neXt Curve research portal at www.next-curve.com for the tech and industry insights that matter.
Mark as Played
Transcript
Episode Transcript
Available transcripts are automatically generated. Complete accuracy is not guaranteed.
Leonard Lee (00:09):
Hey everybody, this is Minor Lee, executive Analyst
at Ncur, and welcome to thisepisode of the Ncur Rethink
Podcast.
It is our Silicon Future seriesthat we do in collaboration with
Cambrian AI research with CarlFloyd and uh, Dave Ulta via of
(00:32):
hot tech, hot tech,
David Altavilla (00:33):
vision and
Leonard Lee (00:34):
analysis.
Yeah.
Nice.
A new edition for this episode.
And of course the missing.
Missing Jim.
The missing Jim
David Altavilla (00:41):
McGregor.
Leonard Lee (00:41):
McGregor
David Altavilla (00:43):
Research, which means we can talk about him.
I know
Karl Freund (00:47):
we're in a bar, maybe he's drinking.
Leonard Lee (00:48):
Yeah, we can say
David Altavilla (00:49):
bad things about him.
He's not here.
Leonard Lee (00:51):
But, before we get started, remember to like, share
and subscribe to Next CurveBriefing, podcast.
And also later on we'll provideyou with some information to,
connect with my colleagues here.
But.
Please remember that, the,comments as well as opinions of
my colleagues here are entirelytheir own and of their own
(01:13):
respective firms and notreflective with next curve.
And,
David Altavilla (01:17):
uh, I don't usually agree with.
Agree with him anyway.
Yeah, I
Leonard Lee (01:20):
never agree with him.
We do that because I don't agreewith these dudes at all.
You know, it's ridiculous.
But, um, we are here at IntelVision in Vision 2025 in Las
Vegas, taking place here at theVirgin Hotel.
And we are in a bar and we areabout to share our hot takes.
(01:44):
Not hot takes actually semicold.
It's like warm maybe.
Yeah.
David Altavilla (01:47):
Yeah.
Leonard Lee (01:47):
From the event.
And this is the first one
Karl Freund (01:51):
with Lapu featuring
Leonard Lee (01:52):
Yeah.
Amp, newly CEO.
Yeah.
David Altavilla (01:57):
Labutta.
Yep.
A new CEO of Intel taking overfor Bat Gelsinger.
Leonard Lee (02:01):
Yes.
And so it's actually a prettybig deal.
It's a big deal.
That one factor just makes thisa very interesting vision for
the company.
Does.
Yeah.
Uh, it's
Karl Freund (02:11):
probably not fair to hold him up to too high
standard.
He's only been on the job for 14days.
Yeah.
David Altavilla (02:17):
He's still drinking from the fire hose.
He's still
Karl Freund (02:19):
drinking from the fire hose.
He, but he laid out a verysimple strategy.
He's gonna focus on culture,he's gonna focus on customers,
he's gonna focus on engineering.
All his keynote was about thosethree things.
Yeah.
but there were hints.
There were hints in there aboutwhat he's might be thinking
about.
I don't think he's made up anykey decisions yet, but he's
(02:39):
getting, he's got his work cutout for him and it's,
David Altavilla (02:41):
yeah, he's passionate about client
computing, of course, whichobviously right now is paying
the bills for Intel.
He's passionate about.
Well, a theme, he talked aboutwas robotics, surprisingly.
So maybe that's a field they'regetting into something new.
But he's also a little bit morepragmatic and measured in his
approach, I think.
And his predecessor, he seems,just a little bit more.
(03:05):
I don't know wanna say reserved,but pragmatic.
Humble, I think is the word.
Humble.
Humble.
He used certainly humble.
Karl Freund (03:10):
He used humble several times.
Yes, he did.
Humility several times.
David Altavilla (03:12):
Humility and humble.
Yeah.
Karl Freund (03:13):
And I think that's an important change in the
culture.
He's gonna try to, he's going toestablish an intel.
I don't think you can.
I was having a conversation withsomebody at today says, you
can't.
Teach people new culture.
He thought, well, that means yougotta replace it.
But he also said he is gonnakeep the key talent.
He's not gonna let the key, it'skey talent go.
And that's very important.
(03:34):
Yeah.
but I suspect there's, he alsomentioned the balance sheet
several times.
Leonard Lee (03:40):
Yeah.
Many, many times.
Karl Freund (03:41):
Well.
The reason you focus on gettingyour balance sheet held healthy
is because you might want to doacquisitions.
Yeah.
Right.
And that's another way to fixthe culture.
You buy the culture you want lipboo.
And his, his investment firm.
Yeah.
He's invested.
In hundreds.
Literally hundreds.
I Googled it before this to makesure I was right.
(04:02):
Uh hundred.
Actually, I didn't Google that.
Perplexity that.
Yeah, if
that's a verb.
I had hundreds of companies thathe's invested in.
So he has a network is trulyimpressive.
Yeah.
So he will have.
Access to both talent and newtechnologies that he could
acquire or mimic, to try to turnthe company around.
(04:22):
The other thing that reallystruck me, I had a conversation
with one of his key executivesthis afternoon and I asked him
how he could contrast hisstrategy with a MD and he said,
we are not.
Going to be a close follower.
Okay?
That means you're either goingto be a loser, which I don't
(04:44):
think he's planning or it meansyou're gonna be a leader.
And so I started askingquestions about what that means
from an AI architecture.
And, uh, he has some veryinteresting ideas I haven't
heard from anybody else.
Yeah.
Um, I'll write about them on my Forbes article
or, okay.
I can share'em here if you want.
Uh, you can go
Leonard Lee (05:02):
ahead and share if you want.
Obviously we want, I don't wantyou to have an open forum Okay.
To share your perspective.
Karl Freund (05:08):
Well, if you think about it, what AMD's doing and
what.
Intel had been doing, yeah.
Is trying to outdo Vidia atbeing Nvidia.
I don't think they do that.
No, you can't.
David Altavilla (05:19):
Not at this point.
not at this point, right.
Yeah.
So too much critical mass there.
Karl Freund (05:24):
We'll see where he takes it.
But instead of just doing bigGPUs with lots of.
Of HBM and really fastnetworking think the question he
was asking is, does the shift tosmaller models change the
infrastructure that's reallyrequired to do world class
inferencing?
And you can imagine all kinds ofthings that they could change.
(05:44):
If he can get the latenciesthroughput, he needs to build,,
world class ai.
David Altavilla (05:49):
Yeah, yeah.
For me,, it's interesting youmentioned the engineering
talent.
He focused on retaining theengineering talent.
That they have now that'sworking, that's getting the job
done.
Enhancing it for sure.
I got that.
And I think as you noted,through his network of
acquisitions and certainly inthe chip industry his tenure at
Cadence brings a lot ofconnections with it too.
(06:12):
Yeah.
Then the other thing for me wasa focus on client.
There was a lot of focus onclient.
There was focus on data centerTwo, they talked about Zion six.
They talked about driving thatroadmap.
But Panther Lake on 18 A isalive and well and strong.
They're very confident.
I was talking with Jim Johnson,of client computing as well on
(06:35):
the side earlier today.
18 A is looking very good.
The defect density at this pointin time, they've shown it, in
little.
Spots, but doing very, verywell.
In fact,
Karl Freund (06:46):
They announced today that it has entered, risk
production.
Right?
Which, which is a big deal for a semiconductor
company.
That means, the chip works, itcan work the power and, you
know, it can deliverperformance.
You've still got some cornercases you need to test for and
so forth, but it means the chipwas ready for early customer
access.
Yep.
And shifts from maybe a coupleof thousand dyes to maybe tens
(07:11):
of thousands of dyes.
Yeah.
Yeah.
That would allow people to testmore at scale.
So that's a huge move for 18 a.
Yeah.
Which the company's pretty muchbet its future on.
David Altavilla (07:20):
Yeah.
Yeah.
No, absolutely.
If Panther Lake executes well,performs well, and 18 a,
performs well.
They're back.
that's gonna be a stake in thesand, at least in client
computing.
They've got, they've got work todo in the data center, which
they know well.
Karl Freund (07:36):
And the work in the data centers been compounded by
their own mistakes, right?
They started out with Multicore.
Yeah.
And then they acquired Nirvanaand then they threw that out
with the bath water and thenthey acquired.
Habana Labs, which one couldargue is it's over.
gouty three actually is beingdeployed now at, at I-B-M-I-B-M
Cloud.
Yeah.
(07:56):
Uh, so IBM chose GOUTY three as their primary
accelerator now in the IBMCloud.
So there's definitely somethingthere that, oh, there's Jim
McGregor.
Yeah,
David Altavilla (08:05):
there's Jim.
He's a little tardy.
Speed.
A little tar.
Jim, are you gonna
Leonard Lee (08:09):
get in here or what?
Tardy man.
Karl Freund (08:11):
We're already
David Altavilla (08:12):
recording.
Dude.
We're already recording.
He's late.
He's late.
Give him a mic.
Give him a mic.
And let him rip.
Let him get in.
Karl Freund (08:20):
Yeah.
So are we doing it without mealready?
You don't need, well
David Altavilla (08:23):
go ahead.
You go.
Time's a wasted dude.
We're not gonna get, yeah,we're.
We're lighting our feet here.
Don't fall though.
Don't knock over the cameras.
Are you guys doing vision onYeah,
yeah, yeah,
yeah.
We're talking vision.
Don't
try to be so sexy.
I'm sure.
Leonard.
Leonard, you cut all this stuff.
Oh, it's the purple shirt.
Jim will just magically appear.
A little bit of both.
(08:44):
Yeah.
Yeah.
So.
Okay.
Of course.
Where was I?
Where was I?
Now we just, you're talkingabout splice it back in.
Oh
Karl Freund (08:51):
yeah.
So now they've canceled thereplacement for HaBO Labs, gouty
three, which was Falcon Chores.
So they've got a lot of work todo, both in terms of hardware
and software, but also in customretention.
Maybe some of, lip boo's, priorbasketball.
Skills will, come in, come toplay as he tries to herd the
customers that are adoptingGouty three.
(09:13):
Yeah.
Yeah.
And getting them to wait andhang on for,
Leonard Lee (09:16):
yeah.
Karl Freund (09:16):
And Jaguar short.
Leonard Lee (09:18):
But I am really glad that you brought, you had
that conversation regarding theai, um, strategy because, I
mean, I'll be.
Quite honest, what I heard up onstage, it sounded more like
bottom feeding approach.
It wasn't very an aggressivestrategy that they were laying
out.
It's more enterprise.
David Altavilla (09:37):
Yeah.
It's, it's inference going afterthe The versus training
application.
Yeah.
Right.
Leonard Lee (09:42):
Right.
But agreed.
You look at it from a bigpicture, it, it wasn't a
compelling, mission that wasputting out Yeah.
It wasn't.
It's like, we'll try to figureout where we fit, but then I was
quite honestly expectingsomething a little bit more bold
from him, given his perspectiveor what we might assume would be
(10:03):
a hate or Yeah.
Have informed and holistic, Ithink.
Yeah.
David Altavilla (10:08):
Situation.
But, I think two weeks in, he'sjust being a little bit, he on
the
Karl Freund (10:14):
board.
Yeah.
Yeah.
Was That's true.
David Altavilla (10:16):
That's true.
A lot.
Yep.
Karl Freund (10:18):
He was on the board let's until last
David Altavilla (10:20):
fall.
You're true, not ignorant.
Very true situation.
I, from
Karl Freund (10:24):
one of his executive, his engineering
executives about what?
Maybe they're thinking about itfrom an engineering standpoint,
and by, by not followingclosely, uh, you're gonna see
some different approaches.
So maybe you don't need biggerGPUs.
Maybe you don't need highbandwidth memory.
Maybe you don't need NVLinkperformance and see that Do
(10:45):
enter.
Leonard Lee (10:46):
That's what I thought was cool to hear that
was, that you had that kind ofconversation and that was what
was Yeah.
Bring up and change questionsbecause I think those are
important questions because oneof the things I heard during
this conference is that AI is,diverse and that connotes hetero
genius.
Yep.
Connotes different types thatneed different types of compute
(11:07):
and so Right.
This whole GPU centric mindsetthat we've been, basically
settled into, because a biggreen guy Yep.
Think, in is dominated the waythat we think about
Karl Freund (11:19):
AI
Leonard Lee (11:20):
computing.
Yep.
Karl Freund (11:20):
So in the last couple years,
David Altavilla (11:23):
the green guy.
The green guy.
Incredible
Karl Freund (11:26):
Hulk.
David Altavilla (11:26):
Yeah,
Karl Freund (11:27):
incredible.
David Altavilla (11:28):
Hu
Leonard Lee (11:28):
was talking.
I
Karl Freund (11:29):
haven't heard Jensen call the credible Hulk.
I've heard other things, but Ihaven't heard him talking.
You know, I'm really afraid
Leonard Lee (11:34):
I'm sitting right next to it.
David Altavilla (11:35):
No, I know.
Something silly.
He'll really get whack.
Yeah.
Yeah.
He might, might stab you orsomething.
Yeah.
Leonard Lee (11:42):
You know, he's like really loose with his, uh, yeah.
And yeah, his fist.
So he's, yeah.
Pointy stuff.
But, uh, yeah.
So what did, what did you think?
I feel like shaking his headwhile I'm talking.
It's like,
Jim McGregor (11:55):
I don't know what you guys have talked about so
far, but I think that, that he'sreiterated the same thing he did
in his, letter to investors,right.
Obviously I think he's got astrategy or what he thinks is
strategy, but he's still talkingand talking to customers,
talking to people within Intel.
A lot of the Intel executiveshaven't even met with him yet.
Yeah.
So he's still making sure that,what's his idea, the right
(12:19):
strategy?
Yeah.
They expect to see that,obviously, focusing on
customers, focusing on, coreproducts.
We'll have to see how thatextends to other markets like
automotive.
Yep.
Some other segments.
He's mentioned robotics.
Yeah.
He's definitely doubling down onFoundry.
Yeah, I think that's reallygood.
And even mentioned, possiblybringing on more talent.
Yeah.
To really, push foundry.
(12:40):
So I don't see a big shift instrategy.
He wants to be moreentrepreneurial, so he's talked
about maybe cutting out later.
So I would expect that, justtrying to reorganize Intel to be
a little bit more reactive tolarger conditions.
Yeah.
I don't know that, and this isgonna take time.
We're at Intel vision, but Idon't think we really saw the
vision yet But I think that willcome and I think that's really
(13:04):
what everyone's really hopingYeah.
In the next six months at least.
I,
Karl Freund (13:08):
I agree.
I, but I think that the thingsyou just mentioned and the
things that we've heard here.
All feel very incremental to me.
Jim McGregor (13:15):
Yeah.
Karl Freund (13:16):
And, they're the things you would expect a new
CEO to say, after 14 days on thejob.
I get that.
Yeah.
Yeah.
But as you point out, he was,he's on the board.
Well, so it's not like hedoesn't know Intel.
So I think.
I think they've got some stuffup their sleeves they're not
talking about yet.
They've got some they're gonnado, they're gonna surprise
people or frankly they're gonnabecome irrelevant.
Jim McGregor (13:36):
And, and knowing his background, I would expect
that you're going to see someacquisition strategists Yes.
In there.
Yes.
As well as divestitures.
Yes.
So I would expect that you'regoing to see drastically
different intel within the nexttwo degrees.
Yeah.
And
David Altavilla (13:49):
I think you have to be careful when you say
he was on the board because.
Just as you noted Jim, he has tomeet with all the product line
executive VPs.
He has to see exactly how thesausage is currently being made.
And boards don't always knowthat what's,
Karl Freund (14:04):
well,
David Altavilla (14:04):
boards don't always know that, right?
No.
No, they don't.
And yeah.
He's gotta meet, he's gottarationalize, he's gotta, really
just get into the weeds Yeah.
With all the different productline directors and, and bu and
figure out, what the pathforward is.
Mm-hmm.
Um, you know, talking to,talking to, to JJ and client.
Um, they are very excited movingforward.
(14:24):
18 a, an lake, as I mentioned,they're excited about simple
stuff like vPro, single sign onfor vPro, which hasn't really
been a big thing.
You know, it's been one of thosethings where Intel's always led
in the enterprise, but.
Enabling it at the fleet levelwas very complicated and hard to
roll out.
And now it's like they're makingit more push button.
Leonard Lee (14:46):
But you know, it's the simple stuff for these guys
that really count.
Yeah.
I mean, like, we covered this,coming out of, cyber Yeah.
Tech, tour, right?
Yeah.
David Altavilla (14:55):
ITT.
Leonard Lee (14:55):
Yeah.
And when, they went into detailabout Lunar Lake, we said, Hey,
if they execute on this needs,it's gonna be a really good
thing for that're shipping a tonof it.
Right.
And then o obviously PantherLake mean it's a pivotal, it's a
pivotal Yeah.
Generation.
David Altavilla (15:15):
Yeah.
Leonard Lee (15:15):
You know,
David Altavilla (15:16):
on their
Leonard Lee (15:16):
roadmap.
David Altavilla (15:17):
So
Leonard Lee (15:18):
not product, but also the,
David Altavilla (15:20):
for the fab, we were talking about that earlier.
Yeah.
Leonard Lee (15:22):
So let's talk about the fab real
David Altavilla (15:24):
quick.
Leonard Lee (15:24):
What did you guys think?
Because right now I, we stillhave a lot of, rumors out there
that it, it's gonna get spun outor
Jim McGregor (15:33):
they can't.
No
Leonard Lee (15:34):
way.
Jim McGregor (15:34):
No way.
Yeah.
I just wanna kill
David Altavilla (15:36):
this.
Yeah, kill it.
Kill it with fire.
Just so ridiculous.
You
Jim McGregor (15:39):
can't sell it to a foreign because you have the US
and all European government SMCsout kind of Yeah.
Kills that right off the batsout.
Karl Freund (15:47):
I mean, uh, Samsung's out can't sell already
Jim McGregor (15:51):
and I think everyone knows, all of us in the
industry know it's reallyimportant and you can't spend
out until it's making moneyanyway.
Right.
So it's not worth it.
It easy indicated that they'regonna double down on it.
That's really important toIntel.
It's really important to theindustry.
I, yeah, let's kill that rumor.
They're not spinning it off.
Intel is definitely keeping, theFoundry service and they're
(16:12):
definitely doubling down on.
David Altavilla (16:13):
I agree with Jim for once.
I don't know why.
I hate to admit, I
Jim McGregor (16:16):
agree with you.
It doesn't happen often.
David Altavilla (16:21):
It's scary when it does.
It's core to their success.
It's,
Jim McGregor (16:24):
well,
David Altavilla (16:25):
they gotta be chip zilla again.
They gotta, they're trying toget back to that.
Jim McGregor (16:28):
And I think we should even kind of preface that
by saying, listen, they'repat's.
Vision was that Foundry's goingto become a differentiator for
Intel and it still may be as wego forward.
If they execute on the foundry,strategy as they're going to
doing, chips, in their productsand bought in for other people,
(16:48):
this may be critical point.
And when we look two to threeyears out, name they'll up and
say there is no reason.
Instead our challenge.
Yeah.
Yeah.
I was wondering what you guys thought.
Samsung isn't, yeah, Samsungstill has products and balance.
Oh yeah.
David Altavilla (17:00):
And packaging is huge.
And they have serious packagingchops.
Great technology in that area.
Leonard Lee (17:06):
So I was not in the data center.
Ai, ECI, ai.
Yep.
David Altavilla (17:12):
These guys were, yeah.
What did you guys fill me in,guys?
I was in client.
So you guys can talk to datacenter.
Let's listen to these guys.
Yeah.
Karl Freund (17:21):
I was really looking to hear how they're
gonna salvage the transition tobecome an AI provider in the
data.
Center given the cancellation ofFalcon Choice.
And I, I like what I heard, butit's, at this point you gotta
(17:41):
show me silicon, right?
Not power.
What did you like about
what you heard?
What did
They're not trying to be a close follower.
They are really, double down onthe transitions in AI towards
ai.
Towards Edge AI and towards, thereason AI and I think they see
an opportunity there.
(18:02):
None of those things existedwhen GPUs were first embedded,
first applied ai.
Maybe there's a, there's justsomething there.
They can do a better jobaccelerating than their
competitors.
David Altavilla (18:15):
We'll see.
I'll add in that.
Intel also has a very strongsoftware ecosystem for enabling,
the data center, market.
Where Nvidia has obviouslycrushed it with software
platform silicon, the wholeshoot and match.
At least Intel has the softwarechops.
A M D's still playing catch upthere, right?
(18:36):
They have.
They have the software resourcesin place to, to help enable
their silicon a little bit.
In all the acceleration they'recurrently doing on Zion is proof
of that.
Early proof of that I would saystill needs to play out in a
bigger game plan.
But Jim, what do you,
Jim McGregor (18:52):
well, I, from what?
Public.
Yes.
Obviously.
They've got a great, they'reexecuting on the Zion program.
I'm impressed with Zion Six.
I'm excited about, Clearwater.
Yeah.
coming out knowing that, they'regonna be using the same
technology that went into LunarLake, advanced Steve go going
into Z.
(19:13):
Product line and
David Altavilla (19:14):
many cores.
Jim McGregor (19:15):
And they're going, they're continuing on with that
efficiency core and thatperformance core.
And they've got a very goodefficiency core strategy.
obviously, there's a lot ofpressure on jack wire shores
Yeah.
execute, for that strategy.
But I think we're also, startingto see a bigger strategy from
them.
Obviously, some of that'spublic.
(19:35):
Some's not, but we'll have towait and see how they execute on
it.
But they're definitely, they'reout there with the market
understanding that this is adata center play.
It's not a process play.
It's not a GPU play, it's notanything.
You have to look at it as asingle system.
Yeah.
And they're headed that same direction, so Yeah.
(19:56):
And I would never count Intelout of anything.
We've seen them come back fromdepth, when they were behind on
these P four bit processors.
Intel has a lot of the keypieces of the puzzle.
And one thing that peopleforget, Intel still Yeah.
Is one of the biggest investorsin r and d in our industry.
Yeah.
From semiconductor processtechnology, interconnect
technology to, materials,technology, you name it.
(20:19):
And I think they're gonna bringthat all to bear.
And I would say that, they'repart of the way from that.
Strategy that, pat Geling laidout a couple years ago.
They've had some stumbles on theAI side, but I think that we're
still a couple years out beforewe see that bull transition,
that full evolution.
Leonard Lee (20:35):
Yeah.
Yeah, yeah.
So let me ask you this,'cause at least up on stage
during the keynote, they seemedreally apologetic about the six
and where did that vibe comefrom?
Jim McGregor (20:47):
No.
Notice that, I think it was moreapologetic about the previous
generation.
Yeah,
Karl Freund (20:52):
I
David Altavilla (20:52):
think they're, I think they're catching up now
with Zion six.
I think they're pretty
Karl Freund (20:55):
happy with Xon six.
Marcus happy, very happy withZion.
Six.
There's, you know, like we were
Leonard Lee (21:02):
Dion six
Karl Freund (21:02):
closed that closed gap.
Leonard Lee (21:04):
It's considerably, it, it
David Altavilla (21:05):
stems from really odd upon.
Yeah.
It's just, you're sensing thecore density per socket.
Right?
Deficit.
They had a bold.
They're, and they've closed thatgap and memory bandwidth, right?
Because the folks in theexperience,
Leonard Lee (21:20):
they seem pretty happy with this.
So, I don't know, it just threwme off a
David Altavilla (21:23):
little
Leonard Lee (21:23):
bit.
And it's just,
David Altavilla (21:24):
and let's not forget, when you talk about data
center CPUs, right?
Not GPUs, CPUs, A MD is still,minority market share by a fair
margin.
I think AMD's like what?
30?
Which is amazing that they're at35%.
35% kudos.
For that.
Who?
A
md for a md from
pulling that, for chipping all that off.
But still, Intel's got a lot ofinertia, in that space.
Karl Freund (21:47):
But I think, yeah, this, the last five years people
been buying it because themomentum, not because it's a
better product.
Agreed.
And I think agreed.
What I'm hearing from the Intelfolks here,'cause they're very.
Proud of the fact that now theyhave a product that really will
compete head to head with AM
Jim McGregor (22:03):
drive.
Yeah.
So I gotta ask you all, uhoh, ifyou had to pick one phrase from
this conference that you putover and over again, or as it
describes it, what would it be?
'cause I've got
David Altavilla (22:14):
one because phrase I was a power forward.
Jim McGregor (22:18):
That's a great one that from,
David Altavilla (22:19):
lip boot.
Lip boot.
Wouldn't imagine about
Jim McGregor (22:21):
you,
David Altavilla (22:22):
man.
I'm gonna fall flat, go toLeonard and come back to me.
I have to think about this.
Leonard Lee (22:28):
We have to remember that this is a partner customer
conference, right?
Yeah.
Yeah.
Humble.
Humble.
Jim McGregor (22:36):
I don't know.
mine would be, and they used itover and over again.
As realistic to me.
Karl Freund (22:41):
Yeah.
Yes, they did.
David Altavilla (22:43):
Over and over and over.
I was gonna say customer.
Customer focused.
Yeah.
Yeah.
Is was what I took away themajor tagline.
But yeah,
Karl Freund (22:52):
it's been a good show.
It's been a good show.
We've, I've learned a lot.
I have.
David Altavilla (22:57):
And it's interesting to hear Intel's so
humble.
Yes.
You know, so.
Leonard Lee (23:04):
Yeah.
I mean, we have heard it beforethough.
We have.
Yeah.
And it's just, I, I think I, I'dlike to see Intel get their AI
story together in a forwardlooking way.
Yeah.
And also in consideration wherethis may go in terms of ai, we
hear Pat talk a lot about hownext.
(23:25):
Big thing in AI is influence.
What does that mean?
'cause I think a lot of folksjust say it.
Yeah.
They don't really articulatewell, what does that dynamic
actually look like?
And then when it gets out towardthe car edge, quote unquote our
edge, how do the economics playout there?
(23:45):
What are the technologies thatthe TransTech we're seeing in AI
that impact how that, howeverything goes from the data
center and.
Right.
Really talking specificallyabout gen ai and the derivatives
above it go out toward, at theedge.
The edge,
David Altavilla (24:02):
well, all the different because I
Jim McGregor (24:04):
think the entire industry's right.
David Altavilla (24:06):
Yeah.
But if you talk AI at the edge,if there's one tech player in
the field right now that has anadvantage in that regard, it's
intel.
From a software standpoint, openVno has been around forever,
ever.
Yeah.
And it's.
Very well established.
It accelerates a lot of appscurrently,
Leonard Lee (24:23):
and you know what,
David Altavilla (24:24):
you know, they have a really good leg to stand
on there.
Leonard Lee (24:27):
Computer vision still is one of the biggest.
Oh, yeah, yeah,
David Altavilla (24:31):
absolutely.
Yeah.
Application camera feedseverywhere.
Baby
AI broadly,
right?
Yeah.
And that,
that's, that's true.
That you see people
Karl Freund (24:38):
have, people tend to equate AI with generative ai,
generative ai.
Still Nory revenue.
Leonard Lee (24:44):
Say something.
It's fairly fledgling.
Appropriated, yeah.
AI has been appropriated by GenAI now to be, yeah.
And so that's why, I always tryto remind people, the thing that
we're trying to track here iswhat kind of impact is
generative AI having from a fouron out?
Right.
'cause there's all this other AIstuff that's been there that
(25:07):
works.
I'm netting that stuff out.
It's like, look.
Yeah.
Spending all this stuff oninfrastructure.
Yep.
How is that translating intovalue out toward, agent AI too?
Yeah.
End customers in the end market.
Right.
It's one thing to sell a bunchof GPUs and GPUs systems for
model training in these massivedata centers.
(25:30):
It's another to actually be ableto monetize, right?
Because monetization, new
David Altavilla (25:34):
experiences at the edge
Jim McGregor (25:36):
embedded world, everyone is looking at this now
all the way down to the smallestmicrocontrollers.
how do we scale AI down to theedge?
So it is the industry challenge.
Agreed.
Yeah.
Fun.
David Altavilla (25:49):
Stuff.
Fun stuff.
Never dull moment for guys likeus.
So,
Leonard Lee (25:52):
Any final words?
David Altavilla (25:54):
I think Q4 into Q1, Q4 this year into Q1 next
year is gonna get veryinteresting.
I, I also think we're lookingat, that's when 18 A comes
online for real.
So, I'm sorry, go ahead Jim.
Jim McGregor (26:06):
No, I, good point.
I also think that, one key pointthat I would take away from
zipper is that he understandsthe value of culture and.
Sick.
Right.
To change the culture ation.
Yeah.
Yep.
Yeah.
Karl Freund (26:18):
And I think that starts with being humble and
he's a very humble man.
Very, I, the one thing I wouldadd, just a final thought, is
that, lip Boo has one of thebest rolodexes in the technology
program.
Yes, he does.
Especially in the semiconductorindustry.
David Altavilla (26:32):
That's a good way of saying it.
Yeah.
Karl Freund (26:33):
And that.
That's gonna open doors for him.
David Altavilla (26:36):
Yeah.
Karl Freund (26:36):
He's gonna create opportunities for him for
partnering and it couldpotentially create opportunities
for acquisition.
David Altavilla (26:42):
Yeah.
Karl Freund (26:42):
So I, I Lipper's Rolodex may be their, may become
their competitive advantage.
David Altavilla (26:47):
If there was a guy that's gonna fill Pat's
shoes and more, it's Yeah.
Probably him.
Yep.
He's get some good, lineagebehind him.
Leonard Lee (26:56):
IF is not going anywhere.
Right.
David Altavilla (26:59):
18 A.
That is the goal line that you,they want keep an eye on 18 a.
Yeah.
That comes out strong.
Yeah.
Yeah.
Things are gonna turn quick
Leonard Lee (27:09):
and
Jim McGregor (27:09):
Right.
Leonard Lee (27:09):
14 A is right behind you.
David Altavilla (27:11):
Yeah.
Yeah.
Leonard Lee (27:12):
Cool.
So, hey gentlemen, thanks forjumping on.
Thanks
David Altavilla (27:16):
for having us.
Thanks for having us buddy.
Guys,
Leonard Lee (27:17):
welcome.
Good
David Altavilla (27:18):
to see you again.
Yeah.
Nice to have you join us.
Yeah, you're one of the firstones,
Leonard Lee (27:21):
right?
Yeah.
Long, long time ago.
Yeah.
coming out of, Taipei for Yes,that's right.
2024.
Oh boy.
So, hey, thanks everyone fortuning in and we hope you
enjoyed the conversation tosharing here.
Of course, remember to like,subscribe and also comment, let
us know what you think, what youthought of our.
Discussion here and, some of theinsights and, takes that we
(27:44):
shared from Intel, fish in 2025.
And, remember to, contact thesegentlemen here and at their
respective companies.
This is Jim Gregor.
Actually, you didn't evenintroduce Jim McGregor interior
Jim McGregor (28:00):
research that Jim has, terry research.com.
Karl Freund (28:03):
Carl dot.
Fre Cambrian ai.com
David Altavilla (28:07):
and Dave Alta Villa for Hot Tech Vision and
analysis.
Dave a@hottech.com.
Leonard Lee (28:13):
And remember to check out Next
curve@www.next-per.com for thetech and industry insights that
matter.
And we will see you on the nextepisode of Silicon Fus.
And if he didn't like anything,it's his fault.
Yeah, it's awesome actually.
It's James problem.
All right.
Take care.
We'll see you next time.
(28:33):
Bye.
Bye.
Hey everybody, this is Minor Lee, executive Analyst
at Ncur, and welcome to thisepisode of the Ncur Rethink
Podcast.
It is our Silicon Future seriesthat we do in collaboration with
Cambrian AI research with CarlFloyd and uh, Dave Ulta via of
(00:32):
hot tech, hot tech,
David Altavilla (00:33):
vision and
Leonard Lee (00:34):
analysis.
Yeah.
Nice.
A new edition for this episode.
And of course the missing.
Missing Jim.
The missing Jim
David Altavilla (00:41):
McGregor.
Leonard Lee (00:41):
McGregor
David Altavilla (00:43):
Research, which means we can talk about him.
I know
Karl Freund (00:47):
we're in a bar, maybe he's drinking.
Leonard Lee (00:48):
Yeah, we can say
David Altavilla (00:49):
bad things about him.
He's not here.
Leonard Lee (00:51):
But, before we get started, remember to like, share
and subscribe to Next CurveBriefing, podcast.
And also later on we'll provideyou with some information to,
connect with my colleagues here.
But.
Please remember that, the,comments as well as opinions of
my colleagues here are entirelytheir own and of their own
(01:13):
respective firms and notreflective with next curve.
And,
David Altavilla (01:17):
uh, I don't usually agree with.
Agree with him anyway.
Yeah, I
Leonard Lee (01:20):
never agree with him.
We do that because I don't agreewith these dudes at all.
You know, it's ridiculous.
But, um, we are here at IntelVision in Vision 2025 in Las
Vegas, taking place here at theVirgin Hotel.
And we are in a bar and we areabout to share our hot takes.
(01:44):
Not hot takes actually semicold.
It's like warm maybe.
Yeah.
David Altavilla (01:47):
Yeah.
Leonard Lee (01:47):
From the event.
And this is the first one
Karl Freund (01:51):
with Lapu featuring
Leonard Lee (01:52):
Yeah.
Amp, newly CEO.
Yeah.
David Altavilla (01:57):
Labutta.
Yep.
A new CEO of Intel taking overfor Bat Gelsinger.
Leonard Lee (02:01):
Yes.
And so it's actually a prettybig deal.
It's a big deal.
That one factor just makes thisa very interesting vision for
the company.
Does.
Yeah.
Uh, it's
Karl Freund (02:11):
probably not fair to hold him up to too high
standard.
He's only been on the job for 14days.
Yeah.
David Altavilla (02:17):
He's still drinking from the fire hose.
He's still
Karl Freund (02:19):
drinking from the fire hose.
He, but he laid out a verysimple strategy.
He's gonna focus on culture,he's gonna focus on customers,
he's gonna focus on engineering.
All his keynote was about thosethree things.
Yeah.
but there were hints.
There were hints in there aboutwhat he's might be thinking
about.
I don't think he's made up anykey decisions yet, but he's
(02:39):
getting, he's got his work cutout for him and it's,
David Altavilla (02:41):
yeah, he's passionate about client
computing, of course, whichobviously right now is paying
the bills for Intel.
He's passionate about.
Well, a theme, he talked aboutwas robotics, surprisingly.
So maybe that's a field they'regetting into something new.
But he's also a little bit morepragmatic and measured in his
approach, I think.
And his predecessor, he seems,just a little bit more.
(03:05):
I don't know wanna say reserved,but pragmatic.
Humble, I think is the word.
Humble.
Humble.
He used certainly humble.
Karl Freund (03:10):
He used humble several times.
Yes, he did.
Humility several times.
David Altavilla (03:12):
Humility and humble.
Yeah.
Karl Freund (03:13):
And I think that's an important change in the
culture.
He's gonna try to, he's going toestablish an intel.
I don't think you can.
I was having a conversation withsomebody at today says, you
can't.
Teach people new culture.
He thought, well, that means yougotta replace it.
But he also said he is gonnakeep the key talent.
He's not gonna let the key, it'skey talent go.
And that's very important.
(03:34):
Yeah.
but I suspect there's, he alsomentioned the balance sheet
several times.
Leonard Lee (03:40):
Yeah.
Many, many times.
Karl Freund (03:41):
Well.
The reason you focus on gettingyour balance sheet held healthy
is because you might want to doacquisitions.
Yeah.
Right.
And that's another way to fixthe culture.
You buy the culture you want lipboo.
And his, his investment firm.
Yeah.
He's invested.
In hundreds.
Literally hundreds.
I Googled it before this to makesure I was right.
(04:02):
Uh hundred.
Actually, I didn't Google that.
Perplexity that.
Yeah, if
that's a verb.
I had hundreds of companies thathe's invested in.
So he has a network is trulyimpressive.
Yeah.
So he will have.
Access to both talent and newtechnologies that he could
acquire or mimic, to try to turnthe company around.
(04:22):
The other thing that reallystruck me, I had a conversation
with one of his key executivesthis afternoon and I asked him
how he could contrast hisstrategy with a MD and he said,
we are not.
Going to be a close follower.
Okay?
That means you're either goingto be a loser, which I don't
(04:44):
think he's planning or it meansyou're gonna be a leader.
And so I started askingquestions about what that means
from an AI architecture.
And, uh, he has some veryinteresting ideas I haven't
heard from anybody else.
Yeah.
Um, I'll write about them on my Forbes article
or, okay.
I can share'em here if you want.
Uh, you can go
Leonard Lee (05:02):
ahead and share if you want.
Obviously we want, I don't wantyou to have an open forum Okay.
To share your perspective.
Karl Freund (05:08):
Well, if you think about it, what AMD's doing and
what.
Intel had been doing, yeah.
Is trying to outdo Vidia atbeing Nvidia.
I don't think they do that.
No, you can't.
David Altavilla (05:19):
Not at this point.
not at this point, right.
Yeah.
So too much critical mass there.
Karl Freund (05:24):
We'll see where he takes it.
But instead of just doing bigGPUs with lots of.
Of HBM and really fastnetworking think the question he
was asking is, does the shift tosmaller models change the
infrastructure that's reallyrequired to do world class
inferencing?
And you can imagine all kinds ofthings that they could change.
(05:44):
If he can get the latenciesthroughput, he needs to build,,
world class ai.
David Altavilla (05:49):
Yeah, yeah.
For me,, it's interesting youmentioned the engineering
talent.
He focused on retaining theengineering talent.
That they have now that'sworking, that's getting the job
done.
Enhancing it for sure.
I got that.
And I think as you noted,through his network of
acquisitions and certainly inthe chip industry his tenure at
Cadence brings a lot ofconnections with it too.
(06:12):
Yeah.
Then the other thing for me wasa focus on client.
There was a lot of focus onclient.
There was focus on data centerTwo, they talked about Zion six.
They talked about driving thatroadmap.
But Panther Lake on 18 A isalive and well and strong.
They're very confident.
I was talking with Jim Johnson,of client computing as well on
(06:35):
the side earlier today.
18 A is looking very good.
The defect density at this pointin time, they've shown it, in
little.
Spots, but doing very, verywell.
In fact,
Karl Freund (06:46):
They announced today that it has entered, risk
production.
Right?
Which, which is a big deal for a semiconductor
company.
That means, the chip works, itcan work the power and, you
know, it can deliverperformance.
You've still got some cornercases you need to test for and
so forth, but it means the chipwas ready for early customer
access.
Yep.
And shifts from maybe a coupleof thousand dyes to maybe tens
(07:11):
of thousands of dyes.
Yeah.
Yeah.
That would allow people to testmore at scale.
So that's a huge move for 18 a.
Yeah.
Which the company's pretty muchbet its future on.
David Altavilla (07:20):
Yeah.
Yeah.
No, absolutely.
If Panther Lake executes well,performs well, and 18 a,
performs well.
They're back.
that's gonna be a stake in thesand, at least in client
computing.
They've got, they've got work todo in the data center, which
they know well.
Karl Freund (07:36):
And the work in the data centers been compounded by
their own mistakes, right?
They started out with Multicore.
Yeah.
And then they acquired Nirvanaand then they threw that out
with the bath water and thenthey acquired.
Habana Labs, which one couldargue is it's over.
gouty three actually is beingdeployed now at, at I-B-M-I-B-M
Cloud.
Yeah.
(07:56):
Uh, so IBM chose GOUTY three as their primary
accelerator now in the IBMCloud.
So there's definitely somethingthere that, oh, there's Jim
McGregor.
Yeah,
David Altavilla (08:05):
there's Jim.
He's a little tardy.
Speed.
A little tar.
Jim, are you gonna
Leonard Lee (08:09):
get in here or what?
Tardy man.
Karl Freund (08:11):
We're already
David Altavilla (08:12):
recording.
Dude.
We're already recording.
He's late.
He's late.
Give him a mic.
Give him a mic.
And let him rip.
Let him get in.
Karl Freund (08:20):
Yeah.
So are we doing it without mealready?
You don't need, well
David Altavilla (08:23):
go ahead.
You go.
Time's a wasted dude.
We're not gonna get, yeah,we're.
We're lighting our feet here.
Don't fall though.
Don't knock over the cameras.
Are you guys doing vision onYeah,
yeah, yeah,
yeah.
We're talking vision.
Don't
try to be so sexy.
I'm sure.
Leonard.
Leonard, you cut all this stuff.
Oh, it's the purple shirt.
Jim will just magically appear.
A little bit of both.
(08:44):
Yeah.
Yeah.
So.
Okay.
Of course.
Where was I?
Where was I?
Now we just, you're talkingabout splice it back in.
Oh
Karl Freund (08:51):
yeah.
So now they've canceled thereplacement for HaBO Labs, gouty
three, which was Falcon Chores.
So they've got a lot of work todo, both in terms of hardware
and software, but also in customretention.
Maybe some of, lip boo's, priorbasketball.
Skills will, come in, come toplay as he tries to herd the
customers that are adoptingGouty three.
(09:13):
Yeah.
Yeah.
And getting them to wait andhang on for,
Leonard Lee (09:16):
yeah.
Karl Freund (09:16):
And Jaguar short.
Leonard Lee (09:18):
But I am really glad that you brought, you had
that conversation regarding theai, um, strategy because, I
mean, I'll be.
Quite honest, what I heard up onstage, it sounded more like
bottom feeding approach.
It wasn't very an aggressivestrategy that they were laying
out.
It's more enterprise.
David Altavilla (09:37):
Yeah.
It's, it's inference going afterthe The versus training
application.
Yeah.
Right.
Leonard Lee (09:42):
Right.
But agreed.
You look at it from a bigpicture, it, it wasn't a
compelling, mission that wasputting out Yeah.
It wasn't.
It's like, we'll try to figureout where we fit, but then I was
quite honestly expectingsomething a little bit more bold
from him, given his perspectiveor what we might assume would be
(10:03):
a hate or Yeah.
Have informed and holistic, Ithink.
Yeah.
David Altavilla (10:08):
Situation.
But, I think two weeks in, he'sjust being a little bit, he on
the
Karl Freund (10:14):
board.
Yeah.
Yeah.
Was That's true.
David Altavilla (10:16):
That's true.
A lot.
Yep.
Karl Freund (10:18):
He was on the board let's until last
David Altavilla (10:20):
fall.
You're true, not ignorant.
Very true situation.
I, from
Karl Freund (10:24):
one of his executive, his engineering
executives about what?
Maybe they're thinking about itfrom an engineering standpoint,
and by, by not followingclosely, uh, you're gonna see
some different approaches.
So maybe you don't need biggerGPUs.
Maybe you don't need highbandwidth memory.
Maybe you don't need NVLinkperformance and see that Do
(10:45):
enter.
Leonard Lee (10:46):
That's what I thought was cool to hear that
was, that you had that kind ofconversation and that was what
was Yeah.
Bring up and change questionsbecause I think those are
important questions because oneof the things I heard during
this conference is that AI is,diverse and that connotes hetero
genius.
Yep.
Connotes different types thatneed different types of compute
(11:07):
and so Right.
This whole GPU centric mindsetthat we've been, basically
settled into, because a biggreen guy Yep.
Think, in is dominated the waythat we think about
Karl Freund (11:19):
AI
Leonard Lee (11:20):
computing.
Yep.
Karl Freund (11:20):
So in the last couple years,
David Altavilla (11:23):
the green guy.
The green guy.
Incredible
Karl Freund (11:26):
Hulk.
David Altavilla (11:26):
Yeah,
Karl Freund (11:27):
incredible.
David Altavilla (11:28):
Hu
Leonard Lee (11:28):
was talking.
I
Karl Freund (11:29):
haven't heard Jensen call the credible Hulk.
I've heard other things, but Ihaven't heard him talking.
You know, I'm really afraid
Leonard Lee (11:34):
I'm sitting right next to it.
David Altavilla (11:35):
No, I know.
Something silly.
He'll really get whack.
Yeah.
Yeah.
He might, might stab you orsomething.
Yeah.
Leonard Lee (11:42):
You know, he's like really loose with his, uh, yeah.
And yeah, his fist.
So he's, yeah.
Pointy stuff.
But, uh, yeah.
So what did, what did you think?
I feel like shaking his headwhile I'm talking.
It's like,
Jim McGregor (11:55):
I don't know what you guys have talked about so
far, but I think that, that he'sreiterated the same thing he did
in his, letter to investors,right.
Obviously I think he's got astrategy or what he thinks is
strategy, but he's still talkingand talking to customers,
talking to people within Intel.
A lot of the Intel executiveshaven't even met with him yet.
Yeah.
So he's still making sure that,what's his idea, the right
(12:19):
strategy?
Yeah.
They expect to see that,obviously, focusing on
customers, focusing on, coreproducts.
We'll have to see how thatextends to other markets like
automotive.
Yep.
Some other segments.
He's mentioned robotics.
Yeah.
He's definitely doubling down onFoundry.
Yeah, I think that's reallygood.
And even mentioned, possiblybringing on more talent.
Yeah.
To really, push foundry.
(12:40):
So I don't see a big shift instrategy.
He wants to be moreentrepreneurial, so he's talked
about maybe cutting out later.
So I would expect that, justtrying to reorganize Intel to be
a little bit more reactive tolarger conditions.
Yeah.
I don't know that, and this isgonna take time.
We're at Intel vision, but Idon't think we really saw the
vision yet But I think that willcome and I think that's really
(13:04):
what everyone's really hopingYeah.
In the next six months at least.
I,
Karl Freund (13:08):
I agree.
I, but I think that the thingsyou just mentioned and the
things that we've heard here.
All feel very incremental to me.
Jim McGregor (13:15):
Yeah.
Karl Freund (13:16):
And, they're the things you would expect a new
CEO to say, after 14 days on thejob.
I get that.
Yeah.
Yeah.
But as you point out, he was,he's on the board.
Well, so it's not like hedoesn't know Intel.
So I think.
I think they've got some stuffup their sleeves they're not
talking about yet.
They've got some they're gonnado, they're gonna surprise
people or frankly they're gonnabecome irrelevant.
Jim McGregor (13:36):
And, and knowing his background, I would expect
that you're going to see someacquisition strategists Yes.
In there.
Yes.
As well as divestitures.
Yes.
So I would expect that you'regoing to see drastically
different intel within the nexttwo degrees.
Yeah.
And
David Altavilla (13:49):
I think you have to be careful when you say
he was on the board because.
Just as you noted Jim, he has tomeet with all the product line
executive VPs.
He has to see exactly how thesausage is currently being made.
And boards don't always knowthat what's,
Karl Freund (14:04):
well,
David Altavilla (14:04):
boards don't always know that, right?
No.
No, they don't.
And yeah.
He's gotta meet, he's gottarationalize, he's gotta, really
just get into the weeds Yeah.
With all the different productline directors and, and bu and
figure out, what the pathforward is.
Mm-hmm.
Um, you know, talking to,talking to, to JJ and client.
Um, they are very excited movingforward.
(14:24):
18 a, an lake, as I mentioned,they're excited about simple
stuff like vPro, single sign onfor vPro, which hasn't really
been a big thing.
You know, it's been one of thosethings where Intel's always led
in the enterprise, but.
Enabling it at the fleet levelwas very complicated and hard to
roll out.
And now it's like they're makingit more push button.
Leonard Lee (14:46):
But you know, it's the simple stuff for these guys
that really count.
Yeah.
I mean, like, we covered this,coming out of, cyber Yeah.
Tech, tour, right?
Yeah.
David Altavilla (14:55):
ITT.
Leonard Lee (14:55):
Yeah.
And when, they went into detailabout Lunar Lake, we said, Hey,
if they execute on this needs,it's gonna be a really good
thing for that're shipping a tonof it.
Right.
And then o obviously PantherLake mean it's a pivotal, it's a
pivotal Yeah.
Generation.
David Altavilla (15:15):
Yeah.
Leonard Lee (15:15):
You know,
David Altavilla (15:16):
on their
Leonard Lee (15:16):
roadmap.
David Altavilla (15:17):
So
Leonard Lee (15:18):
not product, but also the,
David Altavilla (15:20):
for the fab, we were talking about that earlier.
Yeah.
Leonard Lee (15:22):
So let's talk about the fab real
David Altavilla (15:24):
quick.
Leonard Lee (15:24):
What did you guys think?
Because right now I, we stillhave a lot of, rumors out there
that it, it's gonna get spun outor
Jim McGregor (15:33):
they can't.
No
Leonard Lee (15:34):
way.
Jim McGregor (15:34):
No way.
Yeah.
I just wanna kill
David Altavilla (15:36):
this.
Yeah, kill it.
Kill it with fire.
Just so ridiculous.
You
Jim McGregor (15:39):
can't sell it to a foreign because you have the US
and all European government SMCsout kind of Yeah.
Kills that right off the batsout.
Karl Freund (15:47):
I mean, uh, Samsung's out can't sell already
Jim McGregor (15:51):
and I think everyone knows, all of us in the
industry know it's reallyimportant and you can't spend
out until it's making moneyanyway.
Right.
So it's not worth it.
It easy indicated that they'regonna double down on it.
That's really important toIntel.
It's really important to theindustry.
I, yeah, let's kill that rumor.
They're not spinning it off.
Intel is definitely keeping, theFoundry service and they're
(16:12):
definitely doubling down on.
David Altavilla (16:13):
I agree with Jim for once.
I don't know why.
I hate to admit, I
Jim McGregor (16:16):
agree with you.
It doesn't happen often.
David Altavilla (16:21):
It's scary when it does.
It's core to their success.
It's,
Jim McGregor (16:24):
well,
David Altavilla (16:25):
they gotta be chip zilla again.
They gotta, they're trying toget back to that.
Jim McGregor (16:28):
And I think we should even kind of preface that
by saying, listen, they'repat's.
Vision was that Foundry's goingto become a differentiator for
Intel and it still may be as wego forward.
If they execute on the foundry,strategy as they're going to
doing, chips, in their productsand bought in for other people,
(16:48):
this may be critical point.
And when we look two to threeyears out, name they'll up and
say there is no reason.
Instead our challenge.
Yeah.
Yeah.
I was wondering what you guys thought.
Samsung isn't, yeah, Samsungstill has products and balance.
Oh yeah.
David Altavilla (17:00):
And packaging is huge.
And they have serious packagingchops.
Great technology in that area.
Leonard Lee (17:06):
So I was not in the data center.
Ai, ECI, ai.
Yep.
David Altavilla (17:12):
These guys were, yeah.
What did you guys fill me in,guys?
I was in client.
So you guys can talk to datacenter.
Let's listen to these guys.
Yeah.
Karl Freund (17:21):
I was really looking to hear how they're
gonna salvage the transition tobecome an AI provider in the
data.
Center given the cancellation ofFalcon Choice.
And I, I like what I heard, butit's, at this point you gotta
(17:41):
show me silicon, right?
Not power.
What did you like about
what you heard?
What did
They're not trying to be a close follower.
They are really, double down onthe transitions in AI towards
ai.
Towards Edge AI and towards, thereason AI and I think they see
an opportunity there.
(18:02):
None of those things existedwhen GPUs were first embedded,
first applied ai.
Maybe there's a, there's justsomething there.
They can do a better jobaccelerating than their
competitors.
David Altavilla (18:15):
We'll see.
I'll add in that.
Intel also has a very strongsoftware ecosystem for enabling,
the data center, market.
Where Nvidia has obviouslycrushed it with software
platform silicon, the wholeshoot and match.
At least Intel has the softwarechops.
A M D's still playing catch upthere, right?
(18:36):
They have.
They have the software resourcesin place to, to help enable
their silicon a little bit.
In all the acceleration they'recurrently doing on Zion is proof
of that.
Early proof of that I would saystill needs to play out in a
bigger game plan.
But Jim, what do you,
Jim McGregor (18:52):
well, I, from what?
Public.
Yes.
Obviously.
They've got a great, they'reexecuting on the Zion program.
I'm impressed with Zion Six.
I'm excited about, Clearwater.
Yeah.
coming out knowing that, they'regonna be using the same
technology that went into LunarLake, advanced Steve go going
into Z.
(19:13):
Product line and
David Altavilla (19:14):
many cores.
Jim McGregor (19:15):
And they're going, they're continuing on with that
efficiency core and thatperformance core.
And they've got a very goodefficiency core strategy.
obviously, there's a lot ofpressure on jack wire shores
Yeah.
execute, for that strategy.
But I think we're also, startingto see a bigger strategy from
them.
Obviously, some of that'spublic.
(19:35):
Some's not, but we'll have towait and see how they execute on
it.
But they're definitely, they'reout there with the market
understanding that this is adata center play.
It's not a process play.
It's not a GPU play, it's notanything.
You have to look at it as asingle system.
Yeah.
And they're headed that same direction, so Yeah.
(19:56):
And I would never count Intelout of anything.
We've seen them come back fromdepth, when they were behind on
these P four bit processors.
Intel has a lot of the keypieces of the puzzle.
And one thing that peopleforget, Intel still Yeah.
Is one of the biggest investorsin r and d in our industry.
Yeah.
From semiconductor processtechnology, interconnect
technology to, materials,technology, you name it.
(20:19):
And I think they're gonna bringthat all to bear.
And I would say that, they'repart of the way from that.
Strategy that, pat Geling laidout a couple years ago.
They've had some stumbles on theAI side, but I think that we're
still a couple years out beforewe see that bull transition,
that full evolution.
Leonard Lee (20:35):
Yeah.
Yeah, yeah.
So let me ask you this,'cause at least up on stage
during the keynote, they seemedreally apologetic about the six
and where did that vibe comefrom?
Jim McGregor (20:47):
No.
Notice that, I think it was moreapologetic about the previous
generation.
Yeah,
Karl Freund (20:52):
I
David Altavilla (20:52):
think they're, I think they're catching up now
with Zion six.
I think they're pretty
Karl Freund (20:55):
happy with Xon six.
Marcus happy, very happy withZion.
Six.
There's, you know, like we were
Leonard Lee (21:02):
Dion six
Karl Freund (21:02):
closed that closed gap.
Leonard Lee (21:04):
It's considerably, it, it
David Altavilla (21:05):
stems from really odd upon.
Yeah.
It's just, you're sensing thecore density per socket.
Right?
Deficit.
They had a bold.
They're, and they've closed thatgap and memory bandwidth, right?
Because the folks in theexperience,
Leonard Lee (21:20):
they seem pretty happy with this.
So, I don't know, it just threwme off a
David Altavilla (21:23):
little
Leonard Lee (21:23):
bit.
And it's just,
David Altavilla (21:24):
and let's not forget, when you talk about data
center CPUs, right?
Not GPUs, CPUs, A MD is still,minority market share by a fair
margin.
I think AMD's like what?
30?
Which is amazing that they're at35%.
35% kudos.
For that.
Who?
A
md for a md from
pulling that, for chipping all that off.
But still, Intel's got a lot ofinertia, in that space.
Karl Freund (21:47):
But I think, yeah, this, the last five years people
been buying it because themomentum, not because it's a
better product.
Agreed.
And I think agreed.
What I'm hearing from the Intelfolks here,'cause they're very.
Proud of the fact that now theyhave a product that really will
compete head to head with AM
Jim McGregor (22:03):
drive.
Yeah.
So I gotta ask you all, uhoh, ifyou had to pick one phrase from
this conference that you putover and over again, or as it
describes it, what would it be?
'cause I've got
David Altavilla (22:14):
one because phrase I was a power forward.
Jim McGregor (22:18):
That's a great one that from,
David Altavilla (22:19):
lip boot.
Lip boot.
Wouldn't imagine about
Jim McGregor (22:21):
you,
David Altavilla (22:22):
man.
I'm gonna fall flat, go toLeonard and come back to me.
I have to think about this.
Leonard Lee (22:28):
We have to remember that this is a partner customer
conference, right?
Yeah.
Yeah.
Humble.
Humble.
Jim McGregor (22:36):
I don't know.
mine would be, and they used itover and over again.
As realistic to me.
Karl Freund (22:41):
Yeah.
Yes, they did.
David Altavilla (22:43):
Over and over and over.
I was gonna say customer.
Customer focused.
Yeah.
Yeah.
Is was what I took away themajor tagline.
But yeah,
Karl Freund (22:52):
it's been a good show.
It's been a good show.
We've, I've learned a lot.
I have.
David Altavilla (22:57):
And it's interesting to hear Intel's so
humble.
Yes.
You know, so.
Leonard Lee (23:04):
Yeah.
I mean, we have heard it beforethough.
We have.
Yeah.
And it's just, I, I think I, I'dlike to see Intel get their AI
story together in a forwardlooking way.
Yeah.
And also in consideration wherethis may go in terms of ai, we
hear Pat talk a lot about hownext.
(23:25):
Big thing in AI is influence.
What does that mean?
'cause I think a lot of folksjust say it.
Yeah.
They don't really articulatewell, what does that dynamic
actually look like?
And then when it gets out towardthe car edge, quote unquote our
edge, how do the economics playout there?
(23:45):
What are the technologies thatthe TransTech we're seeing in AI
that impact how that, howeverything goes from the data
center and.
Right.
Really talking specificallyabout gen ai and the derivatives
above it go out toward, at theedge.
The edge,
David Altavilla (24:02):
well, all the different because I
Jim McGregor (24:04):
think the entire industry's right.
David Altavilla (24:06):
Yeah.
But if you talk AI at the edge,if there's one tech player in
the field right now that has anadvantage in that regard, it's
intel.
From a software standpoint, openVno has been around forever,
ever.
Yeah.
And it's.
Very well established.
It accelerates a lot of appscurrently,
Leonard Lee (24:23):
and you know what,
David Altavilla (24:24):
you know, they have a really good leg to stand
on there.
Leonard Lee (24:27):
Computer vision still is one of the biggest.
Oh, yeah, yeah,
David Altavilla (24:31):
absolutely.
Yeah.
Application camera feedseverywhere.
Baby
AI broadly,
right?
Yeah.
And that,
that's, that's true.
That you see people
Karl Freund (24:38):
have, people tend to equate AI with generative ai,
generative ai.
Still Nory revenue.
Leonard Lee (24:44):
Say something.
It's fairly fledgling.
Appropriated, yeah.
AI has been appropriated by GenAI now to be, yeah.
And so that's why, I always tryto remind people, the thing that
we're trying to track here iswhat kind of impact is
generative AI having from a fouron out?
Right.
'cause there's all this other AIstuff that's been there that
(25:07):
works.
I'm netting that stuff out.
It's like, look.
Yeah.
Spending all this stuff oninfrastructure.
Yep.
How is that translating intovalue out toward, agent AI too?
Yeah.
End customers in the end market.
Right.
It's one thing to sell a bunchof GPUs and GPUs systems for
model training in these massivedata centers.
(25:30):
It's another to actually be ableto monetize, right?
Because monetization, new
David Altavilla (25:34):
experiences at the edge
Jim McGregor (25:36):
embedded world, everyone is looking at this now
all the way down to the smallestmicrocontrollers.
how do we scale AI down to theedge?
So it is the industry challenge.
Agreed.
Yeah.
Fun.
David Altavilla (25:49):
Stuff.
Fun stuff.
Never dull moment for guys likeus.
So,
Leonard Lee (25:52):
Any final words?
David Altavilla (25:54):
I think Q4 into Q1, Q4 this year into Q1 next
year is gonna get veryinteresting.
I, I also think we're lookingat, that's when 18 A comes
online for real.
So, I'm sorry, go ahead Jim.
Jim McGregor (26:06):
No, I, good point.
I also think that, one key pointthat I would take away from
zipper is that he understandsthe value of culture and.
Sick.
Right.
To change the culture ation.
Yeah.
Yep.
Yeah.
Karl Freund (26:18):
And I think that starts with being humble and
he's a very humble man.
Very, I, the one thing I wouldadd, just a final thought, is
that, lip Boo has one of thebest rolodexes in the technology
program.
Yes, he does.
Especially in the semiconductorindustry.
David Altavilla (26:32):
That's a good way of saying it.
Yeah.
Karl Freund (26:33):
And that.
That's gonna open doors for him.
David Altavilla (26:36):
Yeah.
Karl Freund (26:36):
He's gonna create opportunities for him for
partnering and it couldpotentially create opportunities
for acquisition.
David Altavilla (26:42):
Yeah.
Karl Freund (26:42):
So I, I Lipper's Rolodex may be their, may become
their competitive advantage.
David Altavilla (26:47):
If there was a guy that's gonna fill Pat's
shoes and more, it's Yeah.
Probably him.
Yep.
He's get some good, lineagebehind him.
Leonard Lee (26:56):
IF is not going anywhere.
Right.
David Altavilla (26:59):
18 A.
That is the goal line that you,they want keep an eye on 18 a.
Yeah.
That comes out strong.
Yeah.
Yeah.
Things are gonna turn quick
Leonard Lee (27:09):
and
Jim McGregor (27:09):
Right.
Leonard Lee (27:09):
14 A is right behind you.
David Altavilla (27:11):
Yeah.
Yeah.
Leonard Lee (27:12):
Cool.
So, hey gentlemen, thanks forjumping on.
Thanks
David Altavilla (27:16):
for having us.
Thanks for having us buddy.
Guys,
Leonard Lee (27:17):
welcome.
Good
David Altavilla (27:18):
to see you again.
Yeah.
Nice to have you join us.
Yeah, you're one of the firstones,
Leonard Lee (27:21):
right?
Yeah.
Long, long time ago.
Yeah.
coming out of, Taipei for Yes,that's right.
2024.
Oh boy.
So, hey, thanks everyone fortuning in and we hope you
enjoyed the conversation tosharing here.
Of course, remember to like,subscribe and also comment, let
us know what you think, what youthought of our.
Discussion here and, some of theinsights and, takes that we
(27:44):
shared from Intel, fish in 2025.
And, remember to, contact thesegentlemen here and at their
respective companies.
This is Jim Gregor.
Actually, you didn't evenintroduce Jim McGregor interior
Jim McGregor (28:00):
research that Jim has, terry research.com.
Karl Freund (28:03):
Carl dot.
Fre Cambrian ai.com
David Altavilla (28:07):
and Dave Alta Villa for Hot Tech Vision and
analysis.
Dave a@hottech.com.
Leonard Lee (28:13):
And remember to check out Next
curve@www.next-per.com for thetech and industry insights that
matter.
And we will see you on the nextepisode of Silicon Fus.
And if he didn't like anything,it's his fault.
Yeah, it's awesome actually.
It's James problem.
All right.
Take care.
We'll see you next time.
(28:33):
Bye.
Bye.
Popular Podcasts
On Purpose with Jay Shetty
I’m Jay Shetty host of On Purpose the worlds #1 Mental Health podcast and I’m so grateful you found us. I started this podcast 5 years ago to invite you into conversations and workshops that are designed to help make you happier, healthier and more healed. I believe that when you (yes you) feel seen, heard and understood you’re able to deal with relationship struggles, work challenges and life’s ups and downs with more ease and grace. I interview experts, celebrities, thought leaders and athletes so that we can grow our mindset, build better habits and uncover a side of them we’ve never seen before. New episodes every Monday and Friday. Your support means the world to me and I don’t take it for granted — click the follow button and leave a review to help us spread the love with On Purpose. I can’t wait for you to listen to your first or 500th episode!
24/7 News: The Latest
The latest news in 4 minutes updated every hour, every day.
Crime Junkie
Does hearing about a true crime case always leave you scouring the internet for the truth behind the story? Dive into your next mystery with Crime Junkie. Every Monday, join your host Ashley Flowers as she unravels all the details of infamous and underreported true crime cases with her best friend Brit Prawat. From cold cases to missing persons and heroes in our community who seek justice, Crime Junkie is your destination for theories and stories you won’t hear anywhere else. Whether you're a seasoned true crime enthusiast or new to the genre, you'll find yourself on the edge of your seat awaiting a new episode every Monday. If you can never get enough true crime... Congratulations, you’ve found your people. Follow to join a community of Crime Junkies! Crime Junkie is presented by audiochuck Media Company.