Over the past year, it seems that shares of Taiwan Semiconductor (NYSE:TSM) have been en fuego as investors have anticipated that the company would win Apple's (NASDAQ:AAPL) chip manufacturing business for the iPhone 6. Well, it seems that Apple will indeed be building its next-generation A-chip at Taiwan Semiconductor, but the follow-on chip for the successor to the iPhone 6 looks as though it goes back to Samsung.
How do we know? TSMC said so.
In TSMC's most recent call, the company outright stated that during 2015 it would have meaningfully lower share than a key rival for 14/16-nanometer generation foundry business, although it expects that share to rebound in 2016 and beyond.
Unless Apple decides to use the 20-nanometer manufacturing technology for its A9 chip (very unlikely), or unless Apple doesn't launch the iPhone 6 successor on its traditional cadence (also unlikely), it seems that this business doesn't go to Taiwan Semiconductor -- at least not to begin with.
Samsung and/or GLOBALFOUNDRIES seem to be the winners
If Taiwan Semiconductor isn't building the 2015 iPhone processor, then unless Intel (NASDAQ:INTC) is the stealth supplier of the A9 (which seems unlikely), Samsung, GLOBALFOUNDRIES (using the Samsung-developed 14-nanometer process), or both will supply the chips there.
Interestingly enough, Samsung is Apple's most powerful competitor, and it seems unlikely that Apple would be keen to enable Samsung. However, if Samsung is delivering wafers at more competitive prices and/or on a more competitive timeline, then Apple really has no choice.
It is likely that Apple will want to build as much as it can at GLOBALFOUNDRIES, which is not a direct competitor to Apple, but given that GLOBALFOUNDRIES doesn't quite have the track record that Samsung does, Apple may ultimately end up sourcing from both.
Why isn't Intel making a pass at this?
The big question that's really hard to get an answer for is just why Intel didn't jump in and do everything in its power to be Apple's 14-nanometer manufacturing parters. Intel's reputation for being able to get manufacturing technologies ramped into high yields is legendary, although there is a valid argument that Intel is much less experienced at running a foundry business.
Nevertheless, if Intel were able to get the Apple business, it would immediately grab about 15% of the total mobile applications processor market -- and a much higher percentage of the high end mobile applications processor market. Further, it would deny these volumes to TSMC and Samsung.
This remains a mystery to me.
Foolish bottom line
It seems that Taiwan Semiconductor's time as Apple's preferred foundry may be limited as the Samsung/GLOBALFOUNDRIES tag-team notches back a win here. How much of that volume will come from Samsung and how much from GLOBALFOUNDRIES remains to be seen. Further, Apple may second-source from TSMC once the latter has its 16 FinFET process up and running in high volume.
That said, TSMC has been doing fine without the Apple business all of this time, enjoying mobile growth from the vendors that sell into many Android smartphones, so even if TSMC's share of Apple comes down during late 2015 and into 2016, it's still in a better place than it was in 2013 and earlier.