The only people who will gain from this will be Apple's competition.
What began as a rumor is now news: Apple plans to sever its last tie to Samsung and have its line of ARM processors manufactured by the Taiwan Semiconductor Manufacturing Corp., otherwise known as TSMC.
It will be a while before TSMC actually gets anything from this deal, since Samsung will still be supplying chips to Apple for at least another year. It's a windfall for TSMC and a painful hit for Samsung, no doubt about it. Mark Newman, an analyst at Sanford Bernstein in Hong Kong, estimates that Apple's component orders from Samsung were set to hit around $10 billion last year. Who would want to give that kind of money to their mortal enemy?
Still, Tim, you couldn't make a worse decision. Well you could but I don't want to consider those options.
Intel was a rumored suitor to make the Apple ARM chips and I think Tim Cook will regret not signing with them. Intel is an increasing rarity in semiconductors, a company that makes its own chips. Most are fabless, meaning they don't make their chip, they outsource to a firm like TSMC.
As manufacturing processes have gotten smaller and smaller, it's been harder and harder to stay competitive and some firms have fallen by the wayside, leaving just a few. TSMC is the leader in this field and it has all the big players, which is already proving a problem.
For starters, TSMC has a dubious history when it comes to on-time delivery and making the jump to a new fabrication process. There was a time when every Nvidia earnings call featured CEO Jen-Hsun Huang ranting and complaining about TSMC not being able to provide the company with all the product it needed. It got so bad, Nvidia publicly slammed TSMC at the International Trade Partner Conference, of all places.
The odds are that Apple is smarter than you. Even though Intel's processors are faster, other considerations may be at play. LIke price/performance, programming ease, legacy systems, etc.. I'm disappointed like everyone on this board, but I don't go crazy. Sure ARM stock price is going down, but that doesn't mean that Intel's is going up. Wait for earnings and the conference call. Then we'll have a better idea of what's what.
Well, there are three things that most people are ignoring when it comes to the potential of Apple making a major move to Intel:
1.) The market is changing rapidly, especially with Intel. The speed with which Intel is moving has probably caused Apple to have to totally reconsider its roadmap going forward. I think the Altera switch was really their wake-up call and that wasn't very long ago. It takes a huge amount of work to consider a change and project out the effects.
2.) It takes a very long time to engineer a change once a decision is made- literally years for a move to x86 from ARM. I don't think Intel is going to make ARM chips for the simple reason that they don't need to. They will fill their capacity without Apple. Intel is in the driver's seat because of the clear superiority of their technology and manufacturing.
3.) Even if Apple had made the decision to change, they wouldn't tell anyone until they absolutely had to. We have known for a long time that 14nm was the sweet spot for Apple to make a change. That would mean late 2014 or more likely 2015 for the change. If that were the case, no one would know yet and Apple would still be a very long way from any announcement.
Bottom line: We still don't really know anything about where the Intel-Apple relationship is going except in the short-term. I think the iWatch will be an exception. We should find out if that is an Intel deal in 3-6 months...
They may have signed with tsmc, but that doesn't mean they won't use Intel's chips. I expect Apple to use Bay Trail in one form or other when it comes out. There is a big price gap in between the iPad ($499) and MacBook Air ($999). Rumors of iPhones getting different screen sizes and price points, I expect a Merrifield-based iPhone when it comes out.
They may have signed with tsmc, but that doesn't mean they won't use Intel's chips. I expect Apple to use Bay Trail in one form or other when it comes out. There is a big price gap in between the iPad ($499) and MacBook Air ($999). Rumors of iPhones getting different screen sizes and price points, I expect a Merrifield-based iPhone when it comes out
The only thing to watch out for is if Apple sticks with the 40nm node from TSMC. And if it's truly a year out then the next node Apple may want from TSMC may be 28nm. I have faith that by next year TSMC can finally make 28nm in sizable enough yields to satisfy Apple. Lets be really clear about Apple. Up until last year they were still selling a few computers with Core2Duo processors. That' a processor that dates back to 2006! So, no they don't need the fastest processors, or so they think.
The problem is that these Intel 14nm processors will be all over the market and Apple will once again will be two nodes behind. The iPhone will feel a bit sluggish compared to Intel based Android phones. Battery life will be questionable. If you look at the latest Apple press conference iOS7 looks like they copied Android homework very nicely. Apple seems to be running out of ideas, and they are slow to change hardware.
You are thinking that Apple will have some amazing 20nm processor from TSMC and facing stock outages. That's not a good thing because Apple fans will wait 3 months for their new phone to arrive. I am thinking that in order to avoid the stock outages they will not take the first batches from the newest nodes from TSMC. They will most likely stick with 28nm. That's better for Intel because more consumers would fall off the iPhone bandwagon.
Apple seems to happy to just be a player in the phone market providing just enough interesting stuff to keep consumers somewhat loyal. The lack of leadership from Cook makes Apple just a player instead of a leader in the smartphone segment. Sooner or later consumers will get tired of their small phones and lackluster performance and move on. Android is really the only other game in town.
The only thing to watch out for is if Apple sticks with the 40nm node from TSMC. And if it's truly a year out then the next node Apple may want from TSMC may be 28nm. I have faith that by next year TSMC can finally make 28nm in sizable enough yields to satisfy Apple.
Apple is currently using Samsung's 32nm node, and there is no way they can go 'back' to 40nm. There was talk of the A5 in the latest version of Apple TV being the first (Samsung) 28nm part, but it looks like this is still 32nm, but a complete redesign which resulted in a smaller die.
TSMC doesn't have (general) yield issues on the LP 28nm node, the biggest issue is capacity. I doubt Apple will use TSMC for 28nm.
My bet, you'll see this years Apples A* processors staying with Samsung, but dropping to 28nm.
As for , "iPhone will feel a bit sluggish compared to Intel based Android phones". Apple has a huge advantage in that their environment is native. Basically you are talking native code V Java so Apple could be quite far back on (real) CPU performance but the app will still perform better on their platform (Android's Java stack is poor) . Also, Apples GPU's are far ahead of anyone else which is a big driver for the fluid UI.
Like Nvidia, Qualcomm is also fabless and counts on TSMC to make its chips. And TSMC's inability to deliver has hamstrung Qualcomm, too.
The fact that both companies remains with TSMC and haven't jumped to Globalfoundaries or Intel shows how intertwining these deals can be and how hard they are to extract your company. Apple has wanted out from the Samsung deal for a while and still won't be free until next year.
Intel may be struggling in mobility with the Atom processors, but Intel does yields and manufacturing process migration better than anyone. While TSMC wrestles with 28nm and looking to 20nm, Intel is at 22nm now and moving to 14nm for next year. This is important; the smaller the fabrication design, the less power used.
The stock pickers at Seeking Alpha know this. One columnist pointed out "it is quite possible that Intel might be 2 Moore generations ahead of TSMC and other companies by 2014," a reference to Moore's Law.
Not just that, but TSMC already can't meet the capacity needs of Qualcomm and Nvidia. Now it's going to soak up Apple's demands? Apple has a habit of making huge, all-at-once buys. It's been known to nearly drain the market of DDR or Flash RAM with one buy in preparation for a launch. TSMC can't meet the modest demands of Nvidia and Qualcomm and now it's taking on Apple. Apple is going to rue this decision.
Something will have to give. This might be the impetus for Nvidia to move to Globalfoundaries (I doubt it will ever be able to work with Intel after what Jen-Hsun has said about Intel over the years) and Qualcomm may also abandon ship. In any case, three giant players all demanding capacity from a company that can't satisfy two is a recipe for disaster, or at the very least severe product shortages (which for some people is a disaster).