On Oct. 12, I argued that Intel (NASDAQ:INTC) isn't likely to be hired to build Apple's (NASDAQ:AAPL) A-series of processors anytime soon. Interestingly enough, Intel CEO Brian Krzanich all but confirmed the company isn't likely to manufacture Apple's A-series processors over, at the very least, the next couple of years.
Let's take a closer look at what Krzanich said.
Foundry won't move the needle over the next few years
During the call, an analyst asked Krzanich for an update on the company's progress in becoming a contract chip manufacturer for others. Krzanich indicated that he does see it as a growth opportunity for the company but said it isn't going to "largely move the needle in the next couple of years."
The Apple A-chip manufacturing business is one that's worth billions -- assume roughly 200 million chips per year at between $15 and $20 per chip and it's easily a $3-$4 billion business -- and, for Intel, likely the largest single opportunity out there for contract chip manufacturing work.
If Intel had won the A10 business, then the company would tell investors to expect a material revenue ramp beginning in the first half of next year and extending throughout the year. Similarly, if Intel had won the A11 business, then it is likely that Intel would begin shipping wafers to Apple in the first half of 2017.
In other words, from the statement made by Krzanich, it's clear that Intel isn't going to be getting in on the Apple chip manufacturing business over at least the next two iPhone cycles.
Intel has yet to be win Apple
What's interesting to me is that Intel has yet to win foundry business from Apple despite having arguably the best manufacturing technology available at the time.
For example, Intel went into mass production on its P1271 manufacturing process (22-nanometer node suitable for system-on-chip designs) in 2013 and products were available by September 2013. Had Apple used this process to manufacture its A7 system-on-chip (which was ultimately built by Samsung (OTC:SSNLF) on its 28-nanometer process), the company could have had a faster, lower-power chip than (the frankly already crazy impressive) one that it had already delivered.
For the iPhone 6, Apple probably could have stuck with the 22-nanometer process, which would probably deliver better transistor performance relative to the TSMC (NYSE:TSM) 20-nanometer process that it ultimately used (at the expense of chip density, though).
Finally, Intel started building its first chips on its P1273 process (14-nanometer node suitable for system-on-chip designs) in late 2014; this means that the technology would have been ready in time to build the A9.
But Intel, with a 14-nanometer process that is arguably superior to what the foundries ultimately delivered, still didn't win that business.
Maybe Intel might be able to win business three (or more) years down the line?
We know that Intel hasn't won business at the 14/16-nanometer manufacturing technology generation, and the comments from Krzanich suggest that the company isn't in the loop early on in the 10-nanometer generation either.
While I do think there does exist the possibility that Apple could build its A12 chip on Intel's 10-nanometer process (migrating away from foundry 10-nanometer processes), I suspect that if foundry leader TSMC is successful in getting its 7-nanometer technology into production just a year after it starts 10-nanometer production, this could significantly dampen the appeal of migrating to Intel's 10-nanometer tech for the A12.
Right now, I believe that the foundries are moving at an aggressive enough pace so that Apple may ultimately choose to stick with them, since they are proven foundry partners, rather than jump ship to Intel -- a company with very limited experience ramping very high volume products for external foundry customers.