It doesn't matter how good the iPhone 5 turns out to be: Apple can't sell it if it can't build it. It's not alone: this is the cold equation facing many mobile device manufacturers in the coming months, as supply problems dog the chip manufacturers.
The riskiest products are the highest-performance smartphones and tablets, which drive new chip technologies — and are thus most susceptible to the delays that blunt the cutting edge. It's a rare point of vulnerability for Apple, which relies on massive sales of top-end products and likes very much to maintain absolute control over its supply chain.
Nobody doubts Apple wants to lock down its own chip supplies. It has more options than most. For example, the company has been shopping for semiconductor expertise over the past few years, leading some to suggest it may go as far as setting up its own silicon foundry.
These are insanely expensive: Taiwanese company TSMC has spent nearly $10bn (£6.3bn) on its latest plant. Apple has enough cash to buy anything it likes — but it can't change the laws of semiconductor physics. If it decided to build its own fab from scratch, it would be at least three to five years before it came onstream, and to be competitive, it would have to be cutting edge.
Existing chip companies find that hard enough, even with decades of experience, and the next stage in chip production — sub-20nm geometries on 450mm wafers — is doubly challenging. Starting up under those conditions would be like starting a space programme with a manned mission to Jupiter.
Its existing relationships aren't ideal, either. One major supplier is Samsung, which has to hurt. Other suppliers like TSMC are keenly aware that while Apple is the biggest company in the world, the mobile market is bigger yet, and they aren't prepared to sacrifice other customers in exchange for the Cupertino billions. That's despite it claiming that it has largely fixed its own 28nm fab problems and should be able to meet demand by the end of the year.
Apple's ideal chip partner would be one already leading the pack in fab but one without a major mobile presence
Another option would be to work with the Chinese government, which has already invested heavily in chip fabs and would doubtless be delighted to cut a very favourable deal with Apple in exchange for gaining access to competitive production techniques. That wouldn't solve the problems Apple would face in starting up its own fab, though, and partnerships with authoritarian regimes come with risks all their own.
There are other caveats. The chip industry, and in particular the mobile market, is a vast and thorny thicket of cross-licensing, patent barriers and other IP issues. And, once you've built your own chip fab, you lose a lot of flexibility: it's a lot easier to switch suppliers if you need to change course, than to turn your own supertanker around.
So Apple's ideal chip partner would be one already leading the pack in fab but one without a major mobile presence. It would be one with lots of key intellectual property, and masses of experience in working with the networking, storage and processing techniques that stitch together modern IT. It would be sweeter still if that partner had recently cooled on a long-term relationship with one of Apple's major competitors.
And if that partner was just 10 miles away from Cupertino, just the other side of Ponderosa Park, in — say — Santa Clara... well, that would be just perfect.