Intel To Build Chips For Apple?

I guess the idea of Intel making chips for Apple isn't too far fetched. Intel has the capability and they already make all the processors for Apple computers anyway.

Manufacturing chips on behalf of other companies is a major departure for Intel, which for decades has based its business on using its manufacturing prowess to offer its own PC chips superior to rival products. As PC sales contract and Intel's fabrication plants operate at less than full capacity, the chipmaker sees an opportunity to fill idle production lines while earning new revenue.

Click to expand...

Nope, not gonna happen.

If you hadn't add this part "and they already make all the processors for Apple computers anyway" I'd thought I'm reading a very necro'd thread.

Are they talking about going to something other than x86?

It will definitely be interesting to see how Intel does in the Foundry world ... if they can win the Apple business that would definitely be big ... having a competitive Foundry supplier (like Intel) could definitely shake up the market landscape

GotNoRice said:

Are they talking about going to something other than x86?

Click to expand...

I think Intel is going to be somewhat hesitant to support ARM in any way shape or form, even if it means filling their fab capacity rather than letting it go idle. Intel making ARM SoCs is pretty much waving a white flag and admitting that x86 isn't going to win.

I can see it happening. Apple doesn't want to have to rely on Samsung and Intel seems like a better foundry partner for them if Intel opens up their foundries for more contract work.

pelo said:

I think Intel is going to be somewhat hesitant to support ARM in any way shape or form, even if it means filling their fab capacity rather than letting it go idle. Intel making ARM SoCs is pretty much waving a white flag and admitting that x86 isn't going to win.

Click to expand...

x86 mobile devices might be the best way for Apple to merge the desktop and mobile OSes like they want to.

pelo said:

I think Intel is going to be somewhat hesitant to support ARM in any way shape or form, even if it means filling their fab capacity rather than letting it go idle. Intel making ARM SoCs is pretty much waving a white flag and admitting that x86 isn't going to win.

Click to expand...

Except Intel has lots of idle Fab capacity right now ... the moral highroad is great ... maintaining revenue and profits is better

kirbyrj said:

x86 mobile devices might be the best way for Apple to merge the desktop and mobile OSes like they want to.

Click to expand...

It's going to go the other way rather than x86. In-house ARM SoCs will probably power all future Apple products. Apple already has the best ARM SoCs in the business which they design themselves. The question isn't whether they'll go ARM or x86 for their tablets and smartphones (and even future laptops) - that's already decided. The question is who will fab it? TSMC and GloFo have already invested lots of money towards expansion and increasing capacity. Meanwhile Intel is in war with ARM and fabbing anything ARM would be a sign of defeat, admitting that x86 isn't doing well and that spare capacity is going to ARM SoCs.

It would help Intel financially in the short term, but I think the damage to their image and x86 goals would be catastrophic. I just don't see it happening.

It was about time for Intel to open its fabs. They can't sell everything they are capable of producing these days.

This story is more interesting to me:
http://eetimes.com/electronics-news/4407732/Intel-to-make-14nm-FPGAs-for-Altera

kirbyrj said:

x86 mobile devices might be the best way for Apple to merge the desktop and mobile OSes like they want to.

Click to expand...

This may be Intel's motivation, in general.

Intel has eschewed the ARM fab entirely, and others in the industry have suggested Intel will suffer because of mobile implementations coming in the future.

Intel has pushed smaller and more efficient fab... and they've done it incredibly successfully. With a partner like Apple (and Microsoft, too!) to encourage x86 incursion into mobile Intel can position themselves quite well in that market.

If this is true, and Apple has access to Intel's latest fab nodes, then that will put Apple ahead by over 2 years compared with Samsung, who are still on 32nm.

Intel are close to 14nm manufacturing... That's coming late this year, early next. Could be big news, IF it happens.

kirbyrj said:

x86 mobile devices might be the best way for Apple to merge the desktop and mobile OSes like they want to.

Click to expand...

You're thinking of Microsoft. Apple does not appear to have any interest in merging iOS and OS X. They have shown some interest in bringing iOS features to OS X (and have done so already), but nothing beyond that.

Stimpy88 said:

If this is true, and Apple has access to Intel's latest fab nodes, then that will put Apple ahead by over 2 years compared with Samsung, who are still on 32nm.

Intel are close to 14nm manufacturing... That's coming late this year, early next. Could be big news, IF it happens.

Click to expand...

Samsung, and others, aren't far from 14nm either. I believe that LRCX has finalized their 14nm and working on 10nm (though this requires different etching and substrates).

dgz said:

It was about time for Intel to open its fabs. They can't sell everything they are capable of producing these days.

This story is more interesting to me:
http://eetimes.com/electronics-news/4407732/Intel-to-make-14nm-FPGAs-for-Altera

Click to expand...

Intel has already publicly stated they won't consider fabbing chips for what they call "competitors." Altera isn't a competitor and neither is Cisco (well, not really). ARM is most definitely the competitor. Intel fabbing ARM SoCs would be akin to Intel fabbing AMD64's during the Pentium4 days.

As for why Intel would move into the foundry business, the simple reason is money. Unless Intel can break into the mobile market in a big way, the only real avenue for continued revenue growth is increasing its foundry business. The problem with becoming a foundry, though, is that it gives Intel’s competitors (Qualcomm, AMD, Samsung) access to the very same advanced CMOS processes that have allowed Intel to gain such a dominating lead in the PC market. The profit margins for foundry work are lower, too: Whereas Intel’s gross margins are around 60%, TSMC’s gross margins are around 48%.

According to one analyst, if Intel did manage to bring on Apple as a foundry client, the production of iPad and iPhone SoCs would net Intel an additional $4.2 billion in revenue. $13 billion of capital expenditure doesn’t seem quite so crazy, when you look at it like that. At least as far as the industry analysts are concerned, moving into the foundry business is the only real route for Intel. With the value of its stock dropping 19% over the past year, Wall Street would seem to agree.

Click to expand...

That's from the Extremetech article.

The pricing war is one Intel can't win against ARM unless it decides to open up its fabs to competitors. That also means that x86 isn't going to be dominating any market as those competitors will now have access to the same processes as Intel.

It boils down to $$$. In the last 3 quarters, Intel's stock has dropped by ~30%. The profits have dropped as well and to maintain their margins, Intel has instead opted to let fabs slide idle rather than lowering prices. The fabrication/process business requires monstrous amounts of cash, and given the current market of Intel's x86 space and the poor health of the PC segment, it's unlikely Intel will be able to maintain its lead in the fab space (GloFo and TSMC both have had their revised 20nm BEOL processes taped out late last year).

A company like Apple or Qualcomm doesn't need to worry about investing untold billions a year into fabrication, rather they sign a WSA and pay per wafer. In contrast, Intel can only keep its fabs to itself provided it's still the prettiest gal at the ball and all the boys will throw money at her. That's just not the case anymore.

It could be Apple needs some ultra low power design nodes that intel specializes in.

Imagine an ultra low power entry level iPod with color e-paper screen, and ultra low power signal processor. It's energy use would be negligent. The largest consumer would be the energy dumped into the amplifier for the headphones. You could run for weeks at a time before recharge. Now that would be a catch

pelo said:

That's from the Extremetech article.

The pricing war is one Intel can't win against ARM unless it decides to open up its fabs to competitors. That also means that x86 isn't going to be dominating any market as those competitors will now have access to the same processes as Intel.

It boils down to $$$. In the last 3 quarters, Intel's stock has dropped by ~30%. The profits have dropped as well and to maintain their margins, Intel has instead opted to let fabs slide idle rather than lowering prices. The fabrication/process business requires monstrous amounts of cash, and given the current market of Intel's x86 space and the poor health of the PC segment, it's unlikely Intel will be able to maintain its lead in the fab space (GloFo and TSMC both have had their revised 20nm BEOL processes taped out late last year).

A company like Apple or Qualcomm doesn't need to worry about investing untold billions a year into fabrication, rather they sign a WSA and pay per wafer. In contrast, Intel can only keep its fabs to itself provided it's still the prettiest gal at the ball and all the boys will throw money at her. That's just not the case anymore.

Click to expand...

Intels profits dropping are not indicative of the companies long term health. This includes the mobile market.

What it is indicative of is:
1) Decreased consumer spending on new PC's. The recession is still hitting of a number of industries.

2) Let's face it, to 99% of the consuming public there is no real change from Sandy Bridge to Ivy Bridge. So there isn't a need to upgrade. So intel really currently isn't pushing anything innovative out the door in the desktop PC sector for the last couple of years. Haswell will be their first major redesign since Sandy Bridge.

3) Intel has a massive war chest of funds at it's disposal. They have the best designers, and best process nodes in the industry. When intel realized they were losing to AMD's Athlon they dipped into that warchest once and dropped the P4 architecture and went with new Core architecture. We now see the fruits of that investment today.

Intel is currently dipping into that warchest again to make it's presence known in the mobile market. Intel is investing heaviily into architectures like Atom to make their presence known there. This includes out of order execution Atom processors that consume <2 Watts and still have a greater IPC then ARM.

Another problem isn't so much design envelopes which can beat ARM, it's getting a foothold into the market with CISC-VLSW without alienating current customers running ARM. (Similar to the Apple switchover from Power PC to x86)

So the long and short of it is this:
Intel has excess production ready
Intel has excellent fab nodes
It's a win win to sell out plant time.

The shareholders are concerned though, and perhaps rightfully so.

Intel's Atoms are lackluster. Let's face it, they kinda stink. Clover Trail is a weak tablet SoC that gets trumped by cheaper ARM SoCs (and outperformed as well) while Medfield is a decent phone SoC without LTE.

The problem here is that even if Intel makes awesome Atom SoCs for both tablets and phones (they don't, atm), they can't survive off of those margins. Whereas the ARM players make a killing on small margins and relying on other fabs, Intel doesn't have that option. In order to keep the big blue boat running, Intel requires margins of ~60% and costly chips. In the smartphone and tablet space, a costly processor isn't going to fly with anybody. OEMs can't be bullied into buying a $200 x86 CPU (+$10 chipset) to stick in a tablet when they can get an ARM SoC with very good performance (enough for most people) at a fifth of the cost.

As soon as the PC space shrinks and upgrades are postponed in favor of other devices, Intel faces a steep uphill battle. They're not structured like a Samsung. Intel lacks that flexibility and relies very heavily on OEMs and others supplying critical components. If people start to opt out of that $750 laptop and instead go with the $400-$600 tablet, what's Intel going to offer them? A crappy Clover Trail SoC with the performance of an old netbook that runs Win8? A WiFi card? an expensive SSD? These things aren't going to keep Intel afloat.

I think it's feasible for one reason: Intel will likely have massive excess manufacturing capacity in the near future. With the early move to 450mm wafers, multiple fabs on the leading edge and process node lead, keeping the fabs busy could mean taking on non-traditional (for Intel) foundry-type business. While it may not taking on many customers like traditional foundries do, a very limited number of high volume customers could minimize issues of porting designs to Intel's manufacturing process.

It may also be that Intel had hoped to capture some volume of the smart phone/CE market with Atom to fill that capacity, but that hasn't happened (and I think Intel will struggle with it for a very long time). Intel just retains too much control over processor, even when it offers to allow custom blocks to be added. That and the cost premium over same segment ARM chips makes it too inflexible and expensive for handhelds and tablets outside of tiny niches. But excess capacity is still excess and whatever it takes to fill it, at the right price, makes it a possibility.

Throwing in a little conspiracy, Intel still has an ARM license. While pride may get in the way, an Intel designed (or acquisition of a custom designing an) ARM processor may be in the skunk works, Yamhill-style. It's not a bet, but a hedge that having a performance product as a back up is better than not having one The con side of this is despite the benefits Intel may offer (more advanced manufacturing/higher performance), more expensive chips price themselves out of many designs. It simply may not be worth it with so much competition.

* replace conspiracy with conjecture in the last paragraph

i cant imagine profit margins for intel are any good on the mobile side of things. The benchmarks ive seen from the mobile intel were good, but nothing mind blowing.

buttons said:

i cant imagine profit margins for intel are any good on the mobile side of things. The benchmarks ive seen from the mobile intel were good, but nothing mind blowing.

Click to expand...

Intel's everyday application performance per watt is very very good for Atom.

The problem comes from graphics, a traditional weakpoint for intel.

The long term plan for Atom is 2W/core out of order execution with Haswell graphics attributes.

That my friend is a tall order and why we are taking so long in seeing this product.

To me, honestly, the thought that Intel may consider fabricating Apple Ax processors sounds like hell freezing over. It's an ARM core. Intel has considered ARM-based manufacturers such as Qualcomm as competitors, not AMD.

Is this a sign of Intel bowing down to ARM and beginning to use it to compete others? Unlikely not. This sounds like a purely business decision. Imagine the situation right now. Apple won a shaky $1 billion US lawsuit against Samsung, who is Apple's primary Ax fabricator. Apple uses Intel x86 chips in desktop PCs only. Looking for another fabricator that's not Samsung and can provide both the yield, technology, and the price for Apple means settling for Intel, GloFo, or TSMC. It looks like Apple settled with Intel given the company is already manufacturing x86 processors for their laptops and desktop PCs.

Intel on the other hand is merely offering fabricating space to Apple who are now designing their own ARM-based Ax processors. It's like that local storage business renting out space to people at certain prices depending on the size of storage unit. It is not Intel bowing down to ARM where they'll start making ARM-based Intel mobile choices which would actually be interesting-- ARM-based cores plus Intel GT1/2/3 graphics cores. But, that won't happen in a blue moon.

Why?

Intel is working their damndest to get a power hunger x86 core down to the level of ARM-based processors. We're talking about sub-2W TDP range, where some ARM chips are even less than 1W. It is a friggin' tall order for Intel to do unless they do some serious corner cutting to increase power efficiency of an x86 OOo and in-order core(s) like their current and planned future mobile Atoms.

This is how I see the whole Apple and Intel arrangement.

Apple: "Since we're on contract with you to supply us with Intel Core i3, i5, i7 chips current and future, how about we expand the contract to have you fabricate our current and future Ax processors?"

*Apple waves a few more million dollars in front of Intel*
*Intel looks at their currently half-capacity fabricators losing money due to the decrease of desktop PC sales because of mobile devices*

Intel: "Deal."

Simply put.

Freeze dried Apple chips?