Forget the White House Sideshow. Intel Must Decide What It Wants to Be.

[MKWVentures]

I haven't seen the article but not fabbing chips for the iPhone around 2007 has turned TSMC into a Goliath that is now confronting LBT.

so when Intel decided not to do iphone chips [correctly], Who did Apple choose for foundry?

But Forget the past,

LBT will make intel efficient, improve customer relations. that alone will make Intel successful. Intel has smart people who have great ideas, Intel has IP. intel has market share in CPUS. Fixing efficiency and customer relations will put stock at 50+ again.

 

[KevinK]

so when Intel Intel decided not to do iphone chips [correctly], Who did Apple choose for foundry?

We kind of forget don’t we - the first few iPhone app processors were just modified Samsung app processors, until the Apple-designed A4 came out in 2010. That was done via Samsung foundry, and Samsung continued to be the main foundry until the A8 and A9, shifted things to TSMC, starting in 2014 (production).

TSMC’s big win came from the rest of the mobile ecosystem first (baseband processors and non-Samsung app processors).

 

[tomatoma]

We kind of forget don’t we - the first few iPhone app processors were just modified Samsung app processors, until the Apple-designed A4 came out in 2010. That was done via Samsung foundry, and Samsung continued to be the main foundry until the A8 and A9, shifted things to TSMC, starting in 2014 (production).

TSMC’s big win came from the rest of the mobile ecosystem first (baseband processors and non-Samsung app processors).

If you think about it, it is true
Apple wasn't making SoCs in-house from the beginning.
There is a saying that Intel often missed manufacturing, but there are places where the timeline doesn't match, and it feels a little suspicious.
Well, it's wiser to think about what you can do now and look at the present and the future rather than regretting the past.

 

[tomatoma]

You hearing this from design or foundry or both?

Because in my experience “other side is not collaborating” is often an excuse for failure with collaboration being a two way street.

In other words, is it pushing responsibility?

 

[MKWVentures]

We kind of forget don’t we - the first few iPhone app processors were just modified Samsung app processors, until the Apple-designed A4 came out in 2010. That was done via Samsung foundry, and Samsung continued to be the main foundry until the A8 and A9, shifted things to TSMC, starting in 2014 (production).

TSMC’s big win came from the rest of the mobile ecosystem first (baseband processors and non-Samsung app processors).

Exactly!!!

 

[Daniel Nenni]

You hearing this from design or foundry or both?

Because in my experience “other side is not collaborating” is often an excuse for failure with collaboration being a two way street.

Both sides. They both seemed to want to be rid of each other. This is from former employees so that may be a factor. It seemed like communication was the big problem and that can spiral out of control. Things were just not getting done which is what Lip-Bu has said. Is it fixed? Let's hope so!

I remember hearing about this when Swan was CEO. He signed the TSMC outsourcing deal as a challenge to get manufacturing in line with design. I'm not sure it worked out the way he had envisioned now that Intel is expanding their relationship with TSMC (N2).

 

[user nl]

.........................................

I remember hearing about this when Swan was CEO. He signed the TSMC outsourcing deal as a challenge to get manufacturing in line with design. I'm not sure it worked out the way he had envisioned now that Intel is expanding their relationship with TSMC (N2).

Aligning top-management is toughest and most important cookie there is (blackmail never works in the long run). Especially now when they are already treated as two separate entities by the inside top-boss (first PG and now LBT) / management AND the outside world for a couple of years.

Design is competing with AMD, NVIDIA, BROADCOM, Apple and the rest of the fabless designers and Design has no influence on the operation and success of Foundry. Foundry is competing with TSMC and Samsung, and has no influence on the operation/success of Design.

So, there is no "emotional reason" to keep both sides connected. Even worse, the profit of Design is used to compensate for the losses/capex of Foundry, probably for many years to come. This could easily create resentment and "despair" at Design, as they may feel a complete loss of strategic control. Good competent people hate loss of direct control about their (professional) future, so they eventually leave and look for conditions/environment where they sense again control about their (professional) future. Less competent people stay and both sides, Design and Foundry, simply become less competent over the years.

Tough cookie for LBT, INTEL has a disjoint cultural alignment it seems to me. Unfixable besides a real break-up?

https://companyculture.com/about/

 

[Xebec]

Exactly!!!

The 2005-2008 timeline showed a lot of bad decisions related to mobile, both for Intel and AMD.

- AMD divested Radeon mobile to Qualcomm = "Adreno", exiting the GPU IP for mobile devices that were lucrative for a while

- Intel spent money developing StrongARM / XScale then threw that away to focus on x86 everywhere, and then x86 never received proper investment to scale down to mobile power envelopes that it needed

- Then when they said no to Apple -- Apple became a "not Intel" customer, with further snowball effects of Intel losing x86 for the Mac some years later. Samsung and TSMC then became the two leading edge foundries, and eventually with deeper pockets than Intel.

IIRC even Otellini himself said he didn't realize how much volume there would later be on mobile. That was the miscalculation -- they probably assumed 10-100x less chips than actually became 'desired' by the market in the next 5-10 years.

If they said yes to Apple and were still building fabs for Apple chips, they would still be able to afford dropping the ball on AI, GPUs, and x86 architectures like they are now as TSMC's would likely be (IMO) a full node behind where they are at now and with a lot less revenue to show for future nodes. AI has only recently blown up (3-5 years) as a foundry revenue source).

 

[MKWVentures]

The 2005-2008 timeline showed a lot of bad decisions related to mobile, both for Intel and AMD.

- AMD divested Radeon mobile to Qualcomm = "Adreno", exiting the GPU IP for mobile devices that were lucrative for a while

- Intel spent money developing StrongARM / XScale then threw that away to focus on x86 everywhere, and then x86 never received proper investment to scale down to mobile power envelopes that it needed

- Then when they said no to Apple -- Apple became a "not Intel" customer, with further snowball effects of Intel losing x86 for the Mac some years later. Samsung and TSMC then became the two leading edge foundries, and eventually with deeper pockets than Intel.

IIRC even Otellini himself said he didn't realize how much volume there would later be on mobile. That was the miscalculation -- they probably assumed 10-100x less chips than actually became 'desired' by the market in the next 5-10 years.

If they said yes to Apple and were still building fabs for Apple chips, they would still be able to afford dropping the ball on AI, GPUs, and x86 architectures like they are now as TSMC's would likely be (IMO) a full node behind where they are at now and with a lot less revenue to show for future nodes. AI has only recently blown up (3-5 years) as a foundry revenue source).

we can discuss offline. I was right in the middle of the Xscale, Apple, mobile roadmaps and decisions. Intel spent billions on Arm/XSCALE and product development was too slow to be successful in mobile

On Apple, As I have said before, Intel made the right decision. Intel could not ever meet the cost target and would have lost billions. I know exactly what the target was and exactly what the unit cost would have been. PSO was wrong in his retro assessment because he thought if Intel had scale the costs would come down, they would not have... we looked at that scale effect. There are simple reasons why intel could not meet cost goals based on how Intel processes worked. We can discuss them by Zoom.

Reminder: different companies have different strengths. Intel was good a ramping bleeding edge technologies on product architecture defined by Intel into massive volume by doing whatever it took regardless of cost or headcount. We do not ask Honda to build GPUs or Nvidia to build cars.

Under LBT, after the reorgs, Intel will be much more competitive in todays markets

 

[Xebec]

we can discuss offline. I was right in the middle of the Xscale, Apple, mobile roadmaps and decisions. Intel spent billions on Arm/XSCALE and product development was too slow to be successful in mobile

On Apple, As I have said before, Intel made the right decision. Intel could not ever meet the cost target and would have lost billions. I know exactly what the target was and exactly what the unit cost would have been. PSO was wrong in his retro assessment because he thought if Intel had scale the costs would come down, they would not have... we looked at that scale effect. There are simple reasons why intel could not meet cost goals based on how Intel processes worked. We can discuss them by Zoom.

Reminder: different companies have different strengths. Intel was good a ramping bleeding edge technologies on product architecture defined by Intel into massive volume by doing whatever it took regardless of cost or headcount. We do not ask Honda to build GPUs or Nvidia to build cars.

Under LBT, after the reorgs, Intel will be much more competitive in todays markets

Understand and I would like to learn more; just a few elements I might poke at though -

- Intel later wasted $Bs in 'contra revenue' trying to seed Atom into the mobile market, if they lost $$ to Apple (or just had a margin of .. nothing) would it have been worse than that?

- There is some value in reputation, being Apple's fab partner grants TSMC even more reputation than it would have earned otherwise

- In the long run, if Intel's scale would have run TSMC and later Samsung into the ground (that 2011 chart about 'cost to build leading edge fabs), then Intel could have charged monopoly prices

- If Foundry became successful in volume, even at high cost -- it might have gotten more serious focus on the cost side than Intel fabs have gotten 'in the current timeline'. or at least gotten serious focus earlier than 2025.

- Intel also could have sold some IP or 'leased IP' to Apple if they were fabbing with Intel. There is certainly some IP that would have cost Intel effectively nothing to mark up when partnering with Apple

- The culture shift required for partnering with a "frenemy" might have knocked Intel out of it's arrogance a little earlier. Culture is the core reason they're in the position they're in.

Anyway just peanut gallery stuff from me..

 

[MKWVentures]

Understand and I would like to learn more; just a few elements I might poke at though -

- Intel later wasted $Bs in 'contra revenue' trying to seed Atom into the mobile market, if they lost $$ to Apple (or just had a margin of .. nothing) would it have been worse than that?

- There is some value in reputation, being Apple's fab partner grants TSMC even more reputation than it would have earned otherwise

- In the long run, if Intel's scale would have run TSMC and later Samsung into the ground (that 2011 chart about 'cost to build leading edge fabs), then Intel could have charged monopoly prices

- If Foundry became successful in volume, even at high cost -- it might have gotten more serious focus on the cost side than Intel fabs have gotten 'in the current timeline'. or at least gotten serious focus earlier than 2025.

- Intel also could have sold some IP or 'leased IP' to Apple if they were fabbing with Intel. There is certainly some IP that would have cost Intel effectively nothing to mark up when partnering with Apple

- The culture shift required for partnering with a "frenemy" might have knocked Intel out of it's arrogance a little earlier. Culture is the core reason they're in the position they're in.

Anyway just peanut gallery stuff from me..

please check the timeline. By the time TSMC was working with Apple, they were already on track to be the best executing Fab company in the world.

Contra revenue was hilarious (One of BKs first programs). Intel tried to use its wildly successful rebate program to gain share at Samsung. Rebates were $110 for every 100 spent. Then Samsung dropped them when the price was not longer negative.

PSOs thought was similar to yours.... It would have made Intel more cost effective and less arrogant. Maybe true.... But high cost and arrogance are deeply engrained.... Pats mantra was "Intel exceptionalism" and "18A will be so good, it doesnt matter what we spend... if you build it, they will come"... Pretty much the opposite of Grove and LBT.

I am working on a Zoom presentation on Intel Past and future and challenges. Look for it in a couple weeks on my Website or linked in. you guys can critique me there!

 

[blueone]

Then when they said no to Apple -- Apple became a "not Intel" customer, with further snowball effects of Intel losing x86 for the Mac some years later. Samsung and TSMC then became the two leading edge foundries, and eventually with deeper pockets than Intel.

Intel didn't lose the Mac business because of snowball effects, Intel lost the Mac business because the M-series CPUs are just plain better than Intel client CPUs. I've had both since the M-series was released (my wife has Intel-based laptops for work), and it's not a close comparison.

IIRC even Otellini himself said he didn't realize how much volume there would later be on mobile. That was the miscalculation -- they probably assumed 10-100x less chips than actually became 'desired' by the market in the next 5-10 years.

Higher volume just would have meant more losses. The prices Jobs wanted were not realistic, and I doubt he ever got prices like that from Samsung or TSMC.

If they said yes to Apple and were still building fabs for Apple chips, they would still be able to afford dropping the ball on AI, GPUs, and x86 architectures like they are now as TSMC's would likely be (IMO) a full node behind where they are at now and with a lot less revenue to show for future nodes. AI has only recently blown up (3-5 years) as a foundry revenue source).

This was PSO's real dilemma. Jobs wanted the N-process, which means Intel would have had to build some number of extra fabs. The cost and the commitment bet would have been too long-term and too high. If Jobs was willing to use the N-1 depreciated process that might have been a different story, though I'm still wondering how Intel process of that day supported mobile market power efficiency. I just don't understand the argument that if Intel thought the volume would have been 100x what they were assuming, that it would have changed Intel's outcome. They'd just lose 100x more money.

 

[XtorTourist]

You hearing this from design or foundry or both?

Because in my experience “other side is not collaborating” is often an excuse for failure with collaboration being a two way street.

It's not uncommon for groups to optimize for themselves instead of the system as a whole. I assign much more blame to the management that sits above those groups and is in charge of that system and let the dynamic fester. I think Tan will be much better than previous CEOs in this respect.

 

[Artificer60]

Intel didn't lose the Mac business because of snowball effects, Intel lost the Mac business because the M-series CPUs are just plain better than Intel client CPUs. I've had both since the M-series was released (my wife has Intel-based laptops for work), and it's not a close comparison.

That fact that Intel couldn't deliver on 10nm and had to go down the 14nm+++ road along with major product delays didn't help with this either. My understand is that apple ran out of patience.

 

[tomatoma]

Aligning top-management is toughest and most important cookie there is (blackmail never works in the long run). Especially now when they are already treated as two separate entities by the inside top-boss (first PG and now LBT) / management AND the outside world for a couple of years.

Design is competing with AMD, NVIDIA, BROADCOM, Apple and the rest of the fabless designers and Design has no influence on the operation and success of Foundry. Foundry is competing with TSMC and Samsung, and has no influence on the operation/success of Design.

So, there is no "emotional reason" to keep both sides connected. Even worse, the profit of Design is used to compensate for the losses/capex of Foundry, probably for many years to come. This could easily create resentment and "despair" at Design, as they may feel a complete loss of strategic control. Good competent people hate loss of direct control about their (professional) future, so they eventually leave and look for conditions/environment where they sense again control about their (professional) future. Less competent people stay and both sides, Design and Foundry, simply become less competent over the years.

Tough cookie for LBT, INTEL has a disjoint cultural alignment it seems to me. Unfixable besides a real break-up?

https://companyculture.com/about/

It's not possible

 

[tomatoma]

we can discuss offline. I was right in the middle of the Xscale, Apple, mobile roadmaps and decisions. Intel spent billions on Arm/XSCALE and product development was too slow to be successful in mobile

On Apple, As I have said before, Intel made the right decision. Intel could not ever meet the cost target and would have lost billions. I know exactly what the target was and exactly what the unit cost would have been. PSO was wrong in his retro assessment because he thought if Intel had scale the costs would come down, they would not have... we looked at that scale effect. There are simple reasons why intel could not meet cost goals based on how Intel processes worked. We can discuss them by Zoom.

Reminder: different companies have different strengths. Intel was good a ramping bleeding edge technologies on product architecture defined by Intel into massive volume by doing whatever it took regardless of cost or headcount. We do not ask Honda to build GPUs or Nvidia to build cars.

Under LBT, after the reorgs, Intel will be much more competitive in todays markets

Can reorganization really be competitive?
To be honest, it's dangerous to separate FABs when there are no customers.
Even if it is separated, it will remain in Intel for a certain period of time and should be separated if the customer can secure it.
There is no benefit to being too fast

 

[blueone]

That fact that Intel couldn't deliver on 10nm and had to go down the 14nm+++ road along with major product delays didn't help with this either. My understand is that apple ran out of patience.

The M1 didn't pop out of another universe. Apple was clearly working on it for years. I think they always wanted their own SoCs. I'm most impressed by their using the same M-series CPUs in iPads. My wife's new 13" M3 iPad is pretty amazing.

 

[Xebec]

Intel didn't lose the Mac business because of snowball effects, Intel lost the Mac business because the M-series CPUs are just plain better than Intel client CPUs. I've had both since the M-series was released (my wife has Intel-based laptops for work), and it's not a close comparison.

Higher volume just would have meant more losses. The prices Jobs wanted were not realistic, and I doubt he ever got prices like that from Samsung or TSMC.

This was PSO's real dilemma. Jobs wanted the N-process, which means Intel would have had to build some number of extra fabs. The cost and the commitment bet would have been too long-term and too high. If Jobs was willing to use the N-1 depreciated process that might have been a different story, though I'm still wondering how Intel process of that day supported mobile market power efficiency. I just don't understand the argument that if Intel thought the volume would have been 100x what they were assuming, that it would have changed Intel's outcome. They'd just lose 100x more money.

Unequal answer on the Mac loss; Imtel would still have been fabbing the M1 if they fabbed earlier Apple SoC's.

re: 100X - my understanding (from Intel former employees on this forum) is Intel was always inefficient at fab costs compared to TSMC because of pracrices/discipline and they had little incentive to fix it because the product advantage paid so well. I'm saying that if foundry became a large part of Intel's cost say around 2010, at some point even Intel would have figured out how to fab more cost efficiently.

When a previously small piece of a business becomes a large piece, things tend to change to support that piece..

 

[Xebec]

please check the timeline. By the time TSMC was working with Apple, they were already on track to be the best executing Fab company in the world.

Contra revenue was hilarious (One of BKs first programs). Intel tried to use its wildly successful rebate program to gain share at Samsung. Rebates were $110 for every 100 spent. Then Samsung dropped them when the price was not longer negative.

PSOs thought was similar to yours.... It would have made Intel more cost effective and less arrogant. Maybe true.... But high cost and arrogance are deeply engrained.... Pats mantra was "Intel exceptionalism" and "18A will be so good, it doesnt matter what we spend... if you build it, they will come"... Pretty much the opposite of Grove and LBT.

I am working on a Zoom presentation on Intel Past and future and challenges. Look for it in a couple weeks on my Website or linked in. you guys can critique me there!

re: TSMC, yes agreed - 28nm was spectacular for example, but they were still behind Intel on fab tech for the first half of the 2010s and not significantly ahead until 5nm in 2020. In my fantasy universe they would have been slower going from 28nm to 5nm without Apple as a customer pushing them / writing them large checks.

I always thought Contra was PO, it is even funnier it's BK..

Grove IMO was 'enlightened arrogant' - There's quite a few stories of OEMs (Compaq, etc) that hated dealing with Intel and Andy for arrogance reasons..

look forward to the presentation. I admit you and blueone and others are far more educated than me on this stuff, but I wanted to present some reasons why I think the Apple business loss was the beginning of the end for Intel fabs. I think it became inevitable after that, with the economics requirements of Moore's Law always trending towards 'last man standing'. (whoever has the biggest foundry business will fab the final node).

 

[BruceA]

This Apple moment has come and gone but is really irrelevant in todays situation. But let's go back to that Intel, Paul/Steve moment. Can someone spring back to the days when Paul was offered this and tell me what was Intel's biggest challenge at the time, what was their gross margin, what was their revenue growth. What was the obvious path to 100B revenue and where was Intel's competitive. And at the time given the most optimistic take on the Apple request what it would be?

How Intel Missed the iPhone : The XScale Era

Challenging the conventional narrative on Intel's smartphone miss

thechipletter.substack.com