Will Intel Corporation Finally Build Apple Inc.’s Chips at 10 Nanometers?

Daniel Nenni from SemiWiki said in a recent piece that he has "seen the Intel (INTC 0.97%) Foundry people at Apple (AAPL 0.04%)" and "[has] heard tales of [Intel] aggressively pushing 10nm foundry services."

While only time will tell which company Apple ultimately chooses for its custom processors at the 10-nanometer manufacturing generation, it's worth digging into this claim a bit further.

What timeline are we looking at?
Apple tends to release phones on a one-year cadence and to move to new chip manufacturing technologies with each generation. So, if Apple's A8 in the iPhone 6 and 6 Plus is built on Taiwan Semiconductor's (TSM 2.29%) 20-nanometer process, it follows that next year's A9 will be built on either TSMC's 16-nanometer process or Samsung's (NASDAQOTH: SSNLF) 14-nanometer process.

From there, it's unclear where things go. Both Taiwan Semiconductor and Samsung have announced two different "flavors" of 14/16-nanometer process. The former has a "vanilla" 16-nanometer FinFET and a 16-nanometer FinFET+. Samsung has 14-nanometer LPE, or low power early, and 14-nanometer LPP, or low power performance.

My guess, then, is that Apple will do something like this:

Source: Author conjecture, Chipworks.

What are the odds?
At this point, it is difficult to determine the odds that Intel could win this business. I fully believe Nenni that Intel is "aggressively pushing" 10-nanometer foundry services, but it would be unwise to assume that Taiwan Semiconductor isn't doing the same.

In a more recent SemiWiki piece, Scotten Jones worked out the potential transistor densities of the various foundries' 10-nanometer technologies. His conclusion is that at 10-nanometers, Intel should, as it does at 14-nanometers, have a pretty sizable lead over its competition.

Keep in mind, though, that density and transistor performance aren't the only the only factors that go into such a chip manufacturing agreement. Nevertheless, having access to Intel's manufacturing technology could give Apple a pretty big performance and power edge over the Android vendors.

As far as the odds go, I'm going to play it safe and simply say that I wouldn't factor in a win of Apple's foundry business into any "base case" investment scenario.

Let's go "blue sky" for a moment
Still, I'd be willing to consider such a win as part of a "blue sky" (or, in other words, everything goes right) scenario for Intel.

I'm going to assume Intel can get anywhere from $15 to $20 per hypothetical 10-nanometer A11 chip. From that, assuming Apple ships 110 million iPhone 7s units in the first year of availability, and assuming that Apple also ships 40 million iPads with the Intel-built silicon in its first year of availability, then this could mean anywhere from $2.25 billion to $3 billion in incremental revenue.

If we then assume this business comes in at approximately 50% of gross margin for Intel, that scenario could add anywhere from $1.125 billion to $1.5 billion in gross profit to the chipmaker's bottom line. I would also assume incremental operating expenses at this point would be fairly minimal given that Intel already seems to be spending heavily on foundry-related operations (per the large first quarter losses in its "All Other" segment, which includes foundry operations).

Foolish takeaway
Taiwan Semiconductor, Intel, and the Samsung-Global Foundries partnership will all likely compete aggressively for Apple's foundry business. It will be interesting to see which company wins the battle at the 10-nanometer generation, which could have significant implications for the semiconductor industry as a whole.