Question Intel's future after Pat Gelsinger

[DavidC1]

The whole thing about artificially segmenting Atom to be 1/10 performance is what happens under management trained to be beancounters such as with Paul Otellini.

Back then phone chips were selling $20 so from a beancounter viewpoint the profit margins were really no different than selling real beans - nothing. But as the phone market grew, the ASPs rose and the volume is greater as a whole than the PC market.

Another problem with Intel is that they change their CEOs so often and the strategy has to shift dramatically each time.

As for Android Intel has an Internal AOSP build

Actual application compatibility wasn't that good. You are just talking about the OS. There are many details going in per program that Intel devices did not do so well.

And in 2025 it's correct to say Intel had a build. Since they don't have real Android products, the barrier to entering the market is at least as big if not greater than ARM entering PC space.

I can't see in the most optimistic scenario for Intel how they can grow the market. It's servers, Desktops and Notebooks. Tablets and Phones are out.

 

[511]

And in 2025 it's correct to say Intel had a build. Since they don't have real Android products, the barrier to entering the market is at least as big if not greater than ARM entering PC space.

They are maintaining it cause Chromebook and Android are going to be merged and Chromebook are a good chunk of their client part.

 

[yottabit]

Has there been any verdict yet on if Intel is keeping or killing dGPU?

 

[Win2012R2]

Has there been any verdict yet on if Intel is keeping or killing dGPU?

It seems to me it has already been killed: no chance it will get approved to be fabbed on N3

 

[511]

It seems to me it has already been killed: no chance it will get approved to be fabbed on N3

cause it is getting Fabbed on 18AP

 

[Win2012R2]

cause it is getting Fabbed on 18AP

In 2027?

 

[511]

In 2027?

Yes Xe3P GPU IP is based on Intel's process

 

[Panino Manino]

Oh, I remembered, one of the reasons why Windows Phone failed: "boycotts".
Remember how Goolag and Facebook refused to make apps Youtube and Instagram for it?

 

[Win2012R2]

Yes Xe3P GPU IP is based on Intel's process

It won't be produced

 

[511]

It won't be produced

Why do you think so ? It will be produced and I can bet on that there will be GPUs based on 18AP.

 

[Win2012R2]

I can bet on that there will be GPUs based on 18AP.

Not Intel's dGPUs - unless they magically fix their silicon inefficiency AND obtain 50% margins new CEO wants AND there will be enough capacity to spare.

It's up for the chop for sure, pretty sure decision had already been made. Does not mean they won't use IP for iGPUs, but I reckon soon they will license it from AMD and be done with it.

 

[511]

Not Intel's dGPUs - unless they magically fix their silicon inefficiency AND obtain 50% margins new CEO wants AND there will be enough capacity to spare.

This is underway btw PTL will be the prime example for this both their Display Engine/Media Engine/GPU IP.

It's up for the chop for sure, pretty sure decision had already been made. Does not mean they won't use IP for iGPUs, but I reckon soon they will license it from AMD and be done with it.

This ain't happening re check the earnings call.

 

[Win2012R2]

This is underway btw PTL will be the prime example for this both their Display Engine/Media Engine/GPU IP.

It's not dGPU is it?

This ain't happening re check the earnings call.

Go on, quote me relevant part from the call about dGPUs.

Intel can't compete in dGPU space and they won't allow such waste of silicon.

 

[511]

Go on, quote me relevant part from the call about dGPUs.

Intel can't compete in dGPU space and they won't allow such waste of silicon.

This discipline will improve our execution speed and move up towards a first time bright mindset while also saving development costs. Finally, turning to AI. In the past, we approached AI with a traditional silicon and training centric mindset. Without a cohesive silicon systems software stack and strategy. While we do need to build and consolidate upon our silicon franchise based upon our x86 CPU and our XE GPUs.

https://in.investing.com/news/transcripts/earnings-call-transcript-intels-q2-2025-results-miss-eps-forecast-stock-dips-93CH-4927293

 

[Win2012R2]

While we do need to build and consolidate upon our silicon franchise based upon our x86 CPU and our XE GPUs.

Yeah - "consolidate", pretty openly in corpo speak he says they will cut dGPUs, which is what happens when one "consolidates".

Cutting dGPUs is a no brainer - it's a net loss, nothing gained by doing it, even AMD is getting scraps.

 

[DZero]

Not Intel's dGPUs - unless they magically fix their silicon inefficiency AND obtain 50% margins new CEO wants AND there will be enough capacity to spare.

It's up for the chop for sure, pretty sure decision had already been made. Does not mean they won't use IP for iGPUs, but I reckon soon they will license it from AMD and be done with it.

I don't think so... AMD already has a massive advantage and will get the most benefits from that.
The issue is that x86 depends on Intel mostly. With them weakened, will make ARM and RISC-V stronger.

And Windows will be totally screwed since depends on x86

 

[Win2012R2]

The issue is that x86 depends on Intel mostly

Eh, how?

 

[DZero]

Eh, how?

Most x86 advances are made by Intel. AMD helps, but despite that, is not a really big player. And despite AMD efforts, x86 would be stagnant for a good chunk of time if Intel starts to be screwed up.

Same situation with Windows since depends on x86 the most.

 

[Win2012R2]

Most x86 advances are made by Intel.

 

[511]

Please tell me who wrote the most software for AVX-512 ? Who put resources into open source compilers tool chains ?
Who is taking innovation with AVX 10.2/APX/AMX?
AMD is simply going to copy them and enjoy the benefit of Software just working on their Hardware without having to do much they didn't even manage to get proper compiler support for Zen5 even after it was launched

 

[gdansk]

They didn't even manage to get proper compiler support for Zen5 even after it was launched

🤡Remember what we were discussing months before Zen 5's reveal? GCC diffs for znver5. Oh, you weren't a member then. So maybe not.

But that's hardly on topic here.

 

[511]

🤡Remember what we were discussing before Zen 5's announcement? GCC diffs for znver5. Oh, you weren't a member then. I see.

now i am curious

But that's hardly on topic here.

kind of i was just saying what Intel is doing to forward x86 AMD is just making better hardware and have ample time to figure out better approaches to stuff like AVX-512(anyone remember the OG AVX-512 with the insane heat and clock degradation)

 

[StefanR5R]

Among the x86 innovations with big (if not biggest) positive impact on the ecosystem is Intel® EM64T...

:-)

 

[511]

i have not seen a single EMT64 app in my life

 

[NTMBK]

Most x86 advances are made by Intel. AMD helps, but despite that, is not a really big player. And despite AMD efforts, x86 would be stagnant for a good chunk of time if Intel starts to be screwed up.

Same situation with Windows since depends on x86 the most.

If it wasn't for AMD, Intel would have "innovated" us all into being stuck with Itanium.