> Mobile phone users have welcomed the iPhone. To build such a product, Apple needs to control > all aspects of its form and function down to the design of the Ax SOCs. Delegating this > design to Intel is highly unlikely.
This is a clear "Make-or-Buy" decision.
(a) If/When Apple finds that Intel has made the progression to build great low-power mobile chips; (b) that it (Apple) need not invest huge amount of resources/efforts/dollars to keep up with Intel's SoC and Process roadmaps along with associated risks; and (c) that it is rather onerous and risky to rely on Samsung, TSMC, etc. Then Apple will make the transition to Intel. We don't know where Intel/Apple are in each of the above.
In terms of design uniqueness, Intel has already shown a willingness to build custom SoCs for large partners.
> The last generation iPhone with a 45nm chip is still selling reasonably well today. > What makes you think Apple is desperate for 22nm/14nm. With cost/xsistor going up > because of process complexity, it's unclear how much demand there will be for these > geometries over 32/28. Lower power and higher clock frequencies don't always win over > cost. Higher volume markets like India and China are a lot more cost sensitive.
Selling well today - is all fine and dandy. But this is about the future. Apple has to evolve its products to maintain its image as a cutting edge company.
Customers don't care about 32/28/22/14nm - but they have shown that they love smaller devices that can do more and more. Initial adoption for these smaller geometries (with higher costs) will be by consumers in advanced countries - as volumes increase and costs come down, there will be sufficiently big markets in emerging economies as well.
With each smaller process node and increased wafer size, the complexity and capital costs do go up - but cost per unit comes down drastically. Otherwise, no vendor would even have a roadmap to transition to smaller nodes and larger wafer sizes!
Intel and the chip industry have a track record in innovating to smaller geometries (and larger wafers). I would give them credit that they know what they are doing.
I agree with everything you say. However I also think that Apple will stick with Samsung. What you said makes too much sense and will be the topic of discussion in business schools ten years from now. It will fall under the topic of "bad business decisions."
Though I wouldn't be happy about it, part of me is hoping that Apple pulls off an upset by having purchased every 28nm CPU that TSMC can manufacture. If they had a six month ramp up to build inventory it would put their phone into the zone of interesting.
I am afraid that this won't happen. Apple probably asked around to vendors "What's the node you can make in the tens of millions starting in July?" Samsung said "32nm" and Intel said "32nm for sure, 22nm maybe, we are talking X86, right?" and TSMC said "We are selling wafers that have a few working 28nm on them, it's like gold mining, you never know what you'll get."
Apple will stick with Samsung and run with 32nm for the next two years. They don't perceive hardware as making or breaking their destiny they'll be fine with this decision. In those two years Intel will be in mobile at 22nm and the early stages of 14nm. We're about to find out how unimportant hardware really is.
Whatever decision Apple makes will look like the right one. Apple is going to make the iPhone5 with a larger screen and give iPhone users many of the benefits that Android users now take for granted. Sales will be amazing for this phone, and it will be the "must have" phone of the holiday season. Apple's management will be satisfied.
Over the next two years people will go to phone stores looking to buy the iPhone and walk out with an Android instead. It will happen again and again. With emerging markets where phones cost many weeks salaries they may be a bit more skeptical about spending the extra money for the iPhone.
If Apple doesn't have a phone that sets the bar and establishes them as the market leader based on solid hardware you will see their market share erode as Intel heats up in mobile over the next couple of years. The smartphone market will escape Apple as smartphones become commodities.
Looks like you are taking the position of "Apple continuing with Samsung" since you want to be cautious. Which is conservative and prudent.
However, from time to time, Apple has made big decisions and switches - and they have benefited vastly from these moves. Further, Tim Cook would want to establish himself as a leader by making bold decisions.
If Intel's new Atoms (to be released later this year), process and manufacturing roadmaps are convincing, I would bet that Tim Cook wouldn't hesitate too much to pull the trigger.
Also, though he currently decries hybrid/convertible laptops, I wouldn't be surprised if they have some skunkwork projects inside Apple working on them - of course, these machines would need high performance leading back to Intel.
Doesn't hurt to imagine some sunny skies for Intel :)