News Intel IFS Plans Are a Strategic Mistake, Hints AMD Exec

[Admin]

Time will tell if Intel’s refocusing on manufacturing will pay off, but at least one AMD exec thinks IFS was a mistake.

Intel IFS Plans Are a Strategic Mistake, Hints AMD Exec : Read more

 

[hotaru.hino]

All I can think of is "of course an AMD exec would say this"

 

[rtoaht]

AMD is getting way too cocky with their recent success in data center. But just like Intel recovered and win back the market share in client they could potentially do the same in data center next year. AMD's market share is still way small compared to Intel. They should at least beat Intel in market share before getting this cocky.

 

[TerryLaze]

AMD's business has made strides in recent years, and its fortunes have enjoyed particular growth since it divested in manufacturing. It has also started to look increasingly strong compared to its Goliathan rival, with first-mover advantages in more refined semiconductor processes, pioneering multi-chip packages, and adding potent iGPUs.

 

[JayNor]

I saw a statement that Intel plans to quadruple their advanced packaging capacity. Intel claims that will come into play for IFS customers earlier than the wafer processing.

 

[Kamen Rider Blade]

If Intel wants to make IFS truly succeed, it needs to split up Intel's "Chip Design" side from the Silicon manufacturing side and be truly neutral and compete directly with TSMC and Samsung.

By having their own "Chip Design" side, there is always going to be issues where companies can't fully trust IFS.

Once Intel makes that hard decision, it'll have a reasonable viable road to making IFS succeed, same with Intel's Chip Design side being a Fab-less company.

 

[scottslayer]

Good job on reporting a hint from Intel's direct competitor.

 

[cyrusfox]

Intel is betting the farm on IFS. They have failed to capitalize on opportunities in the past and have only ever been successful in their core business (CPU). They seem to kill off everything else and per the history Darren Grasby words of imminent failure are not unfounded. Only thing different this time is Pat, will he stay on at Intel and push his vision forward to success or will it end in another sell off an further consolidations?

 

[mcmlai]

AMD is getting way to cocky with their recent success in data center. But just like Intel recovered and win back the market share in client they could potentially do the same in data center next year. AMD's market share is still way small compared to Intel. They should at least beat Intel in market share before getting this cocky.

Well, how is he supposed to answer that question? He does have a point. If Intel's 18A fab is ready, are they going to prioritize their own chip? or will they build Apple and Qualcomm chips first? If they prioritize their chip, Apple and Qualcomm are not going to use Intel fab. Just look at Samsung fab. They can't make their own Exynos processor and also fabricate for others.

 

[kjfatl]

Sounds like the opening round of negotiations before AMD begins using IFS as their foundry on some, or all of their designs. If AMD is not willing to seriously consider using IFS, they can expect to pay a 30% premium for TSMC's newer processes.

 

[thestryker]

I'm not sure why anyone would take what this guy is saying seriously outside of it making for salacious headlines. AMD sold off their fab business because the company hemorrhaging money and leading edge fabrication is insanely expensive. Just look at what happened to GlobalFoundries after the spinoff: ditched 7nm development, diversified into custom nodes and didn't start making money for years.

Intel can be successful running a foundry service so long as they're able to deliver capacity. The speed at which they were able to move Intel 4 fabrication from Oregon to Ireland leads me to believe they're preparing for being able to increase capacity rapidly (relatively speaking for fabs). This go around is also the first time Intel has opened up leading edge nodes to third parties which is a big deal and indicator of how seriously they're taking it.

If Intel wants to make IFS truly succeed, it needs to split up Intel's "Chip Design" side from the Silicon manufacturing side and be truly neutral and compete directly with TSMC and Samsung.

By having their own "Chip Design" side, there is always going to be issues where companies can't fully trust IFS.

Once Intel makes that hard decision, it'll have a reasonable viable road to making IFS succeed, same with Intel's Chip Design side being a Fab-less company.

You realize you undermined your already foolish stance by citing Samsung right? 🙄

 

[Kamen Rider Blade]

You realize you undermined your already foolish stance by citing Samsung right?

My bad, should've listed Global Foundary's then.

 

[versed.perception]

Of course Intel wants IFS. With the uptick of datacenter GPU demand FABs are going to be busy through 2028 at this rate, and Intel wants a % of that business.

For competition sake, I really hope Intel can drop a 128core Xeon that can compete with IPC, Clock, and Power per socket up against Epyc before the GPU market starts to saturate due to all this AI crap, until then I will continue to purchase dual Epyc systems going for max density per 1u.

But since they can't seem to do that, they are better off selling their FAB time for any fabless entity that can adapt. It might bring competition to TMSC and drive silicon prices down, which will affect the bottom line driving up costs today. It would be pretty interesting to see Intel FABs working for both AMD and Nvidia as well as Intel's own.

But to keep IFS clean operating Intel should partition their fabs for IFS and subsidize that business. Pay their own way on to the sub to keep it in the competition. But Gelsinger is not the leader that will do this well for Intel, the guy is all over the place and acts like a child when it really counts. I can see him pulling FAB time for Intel's own out from under standing contracts.

 

[bit_user]

Of course this looks odd compared to Intel, because this is what it looks like when you're actually investing in your core business rather than protecting your profits so investors can feast off them. R&D spending can't swing as wildly as revenue. So, when you have a bad quarter or year, you naturally take a big hit on profit.

Now, try looking at AMD's gross revenues, instead.

​

​

 

[thestryker]

Of course this looks odd compared to Intel, because this is what it looks like when you're actually investing in your core business rather than protecting your profits so investors can feast off them. R&D spending can't swing as wildly as revenue. So, when you have a bad quarter or year, you naturally take a big hit on profit.

Now, try looking at AMD's gross revenues, instead.

​

​

While I agree looking at gross revenue is a far better indicator of company success Intel has historically spent approximately the same or more on R&D than AMD makes in revenue period. This has only changed in the last couple of years as AMD's revenue has shot up. So while bad management certainly got Intel to where it is today it's not like they haven't been spending on the core business.

 

[bit_user]

In this case, comparison of Intel vs. AMD's business practices is 100% on-topic. Terry can observe that AMD's profitability hasn't been robust, just as I can counter that AMD is prioritizing growth above profitability. Meanwhile Intel has long been playing a different game of prioritizing returns for investors above investing in their core business. So, of course they look different.

 

[bit_user]

While I agree looking at gross revenue is a far better indicator of company success Intel has historically spent approximately the same or more on R&D than AMD makes in revenue period.

Given Intel's manufacturing operations which AMD lacks, it's a false comparison. If you add TSMC's R&D spending to AMD's, I think the comparison would look entirely different.

BTW, that discrepancy is directly in line with what the AMD exec was talking about.

while bad management certainly got Intel to where it is today it's not like they haven't been spending on the core business.

It's a question of priorities. Intel went over 2 decades of not touching their dividend scheme, and then have been doing considerable share buy-backs on top of that. All the while, doing countless layoffs and divesting from numerous businesses that often seemed like they could still go either way. If your priority is generating returns for investors, then it makes total sense that you protect profitability even if it means underinvesting in the business.

We saw the clear results of this practice with a decade of middling IPC improvements and consumer core-count stagnation, countless corners being cut on security, and repeated failures to deliver on their process roadmap. In my opinion, what's been Intel's biggest challenge isn't its fabs, but rather its investors. A series of CEOs has been unable to meet this challenge, but Pat may turn out to be successful enough. He already cut the dividend and didn't lose his job over it.

 

[TerryLaze]

For competition sake, I really hope Intel can drop a 128core Xeon that can compete with IPC, Clock, and Power per socket up against Epyc before the GPU market starts to saturate due to all this AI crap, until then I will continue to purchase dual Epyc systems going for max density per 1u.

There is no money in that, they can't sell a chip like that for enough money to make it worth their while, that part of the market is drowning in ARM, Intel makes server CPUs for their excellerators and special hardware because that's a monopoly and that's where they make money, also in edge so everything surrounding servers.

Of course this looks odd compared to Intel, because this is what it looks like when you're actually investing in your core business rather than protecting your profits so investors can feast off them. R&D spending can't swing as wildly as revenue. So, when you have a bad quarter or year, you naturally take a big hit on profit.

Now, try looking at AMD's gross revenues, instead.

​

​

What is there to see??????
So they sold their fabs in 2009...when does the revenue increase?! You left out the dates otherwise we would see that the revenue only starts to increase mid 2020. This graph 100% backs up my graph, selling their fabs did diddlysquat to improve their finances, they only started to make money during covid.

 

[Alvar "Miles" Udell]

Selling their fabs and then their HQ in a leaseback arrangement was a way to keep the lights on while they flailed against Intel with a faulty design. They were lucky to eek it out until Zen thanks mostly to people who bought AMD even though they were (far) inferior just because they weren't Intel (ashamedly I am one of them). Then fortune smiled on AMD when Intel ran into troubles at the exact time Zen released so AMD was able to gain the upper hand again.

Intel's IFS appears to be sound at this time as it allows them to generate income and gain fab space until such time as they get their GPU division up to snuff so they have a second stream of income and a way to use that fab space themselves.

 

[bit_user]

What is there to see??????
So they sold their fabs in 2009...when does the revenue increase?!

I think that covered the investment they made in ATI, more than funding new investments. Regardless, it relieved them of a burden that only would've grown over time. It's not obvious they could've invested in Zen, had they continued to be weighed down by their fabs.

You left out the dates otherwise we would see that the revenue only starts to increase mid 2020.

I think the first part of the graph is so flat because their different businesses were probably somewhat out-of-phase, in terms of when they experienced growth. For instance, the 2017-2018 crypto bubble was deflating somewhat, when Zen 2 and Epyc Rome started to ramp. Also, console revenue was tapering off as the Zen-series was getting its legs under it.

they only started to make money during covid.

You're confusing profit with some sense of health or success. For a long time, Amazon famously reinvested quite a lot back into its business, much to the dismay of investors. Just look at them now!

 

[TerryLaze]

I think that covered the investment they made in ATI, more than funding new investments. Regardless, it relieved them of a burden that only would've grown over time. It's not obvious they could've invested in Zen, had they continued to be weighed down by their fabs.

That's great and all, but it's not the topic.
The topic is the exec saying that dropping the fabs allowed them to have growth.
If that growth happens like 15 years later then it's tough to see a correlation between those two...

I think the first part of the graph is so flat because their different businesses were probably somewhat out-of-phase, in terms of when they experienced growth. For instance, the 2017-2018 crypto bubble was deflating somewhat, when Zen 2 and Epyc Rome started to ramp. Also, console revenue was tapering off as the Zen-series was getting its legs under it.

I think it is that flat because amd was trying to sell the exact same CPUs for 10 years and because they had an extremely limited amount of wafers they could turn into lucrative GPUs and because the console deal is apparently making very little money.

You're confusing profit with some sense of health or success. For a long time, Amazon famously reinvested quite a lot back into its business, much to the dismay of investors. Just look at them now!

Wen I say make money while replying to your graph showing revenue then I very obviously mean revenue...
You implied that looking at the revenue would change something but it paints the exact same picture.

 

[Ogotai]

trying to sell the exact same CPUs for 10 years

um, isnt that what intel is pretty much doing ?? even more so during the time they stagnated the mainstream market with quad cores, and were barely given any performance increases year over year ?

 

[bit_user]

That's great and all, but it's not the topic.
The topic is the exec saying that dropping the fabs allowed them to have growth.
If that growth happens like 15 years later then it's tough to see a correlation between those two...

Yes, it's on-topic. Not spending on fab R&D enabled them to focus on their roadmap, like he said. It just took a while for the returns on that investment to accumulate to a sizeable amount, though not 15 years as you incorrectly state.

I think it is that flat because amd was trying to sell the exact same CPUs for 10 years

That's not true. Do not post misinformation.

and because they had an extremely limited amount of wafers they could turn into lucrative GPUs

Source?

because the console deal is apparently making very little money.

Not huge, in any given year, but it's probably one of the things which kept AMD alive from between the launch of the PS4 and Zen.

Wen I say make money while replying to your graph showing revenue then I very obviously mean revenue...

Please say what you mean, then. It's incorrect to state they had zero revenue before covid, and the start of their revenue ramp was actually before covid. They had sequential & YoY growth in each of the last 3 quarters in calendar 2019. It's even more instructive to look at year-on-year growth, as can be seen in this chart:

​

Look at the same data for Intel is much more dire:

​

You implied that looking at the revenue would change something but it paints the exact same picture.

You might not think so, but I think it rounds out the picture. As I said, AMD has been focusing on a growth-oriented strategy and their revenue growth shows they've indeed been successful in that. It's much more apparent what a success Zen has been, if you look at their YoY quarterly growth, where they only had two quarters that were down YoY before 2022 - and those were where we had a dip in the crypto market and console revenues were tapering in the lead-up to PS5.

 

[jkflipflop98]

Someone bookmark this article so in three years when Intel is fabbing AMD chips we can all rub it in his face.

 

[bit_user]

Someone bookmark this article so in three years when Intel is fabbing AMD chips we can all rub it in his face.

I'll say right now I think that could happen. Even so, I've been saying for a while that Intel will probably spin off IFS.

I'm not agreeing with that AMD exec's statement, just taking issue @TerryLaze ' interpretation of AMD's financials.