Well, that didn’t take long. Less than two weeks after Intel announced it would fab chips for LG as part of a new deal to build ARM products directly at Intel, there are rumors that Intel and Apple have sat down to talk about a similar deal between Cupertino and Santa Clara. Reports indicate that new chips could come as early as 2018, though we’d take that with a grain of salt — even when companies eventually shift suppliers, as Apple did by moving to Samsung from TSMC, it can take years just to iron out the contracts.
The Nikkei Asian Review reports that Apple is already trying to nick Apple from TSMC, for all the obvious reasons. Pulling in Apple’s business would give Intel’s custom foundry division instant credibility and establish it as a force to be reckoned with. “TSMC could face tough competition as soon as 2018 or 2019 as Intel is likely to gain orders from Apple by then,” Samuel Wang, a veteran semiconductor analyst at research company Gartner, told the Nikkei Asian Review. “Intel has begun to engage with Apple and it aims to grab one or two top-tier customers from TSMC.”
Intel, TSMC, and Samsung feature sizes at 14nm. Intel leads the market, but can it translate that leadership to advantages for its customers?
Whether or not Intel can grab those customers successfully will depend on whether or not it can adapt its foundry technology and practices to service a very different set of clients. As we covered when Intel canceled its smartphone and tablet chips, Intel already took a serious go at this. Intel’s cost structures weren’t designed to serve the same markets where TSMC, Samsung, and GlobalFoundries find their customers, and the company had difficulty finding a business model for Atom that actually earned it enough money to make competing in the segment profitable.
As for why Apple would want to forge a relationship with Intel, Chipzilla’s foundry knowledge may be to valuable to resist. While TSMC and Samsung have continued to push their own roadmaps, their upcoming 10nm products will apparently be hybridized designs — we’ve heard rumors of 10/20nm, though other sources have stuck to 10/14nm. Either way, Intel is likely to maintain a lead in terms of absolute foundry technology — they’ve still got the only full 14nm die shrink, and they’ll have the only 10nm shrink as well.
The big question is, can Intel turn its foundry prowess into income for itself and its various customers? Apple is undoubtedly interested in Santa Clara’s foundry technology, it’d be a fool not to explore the possibility — but Intel will have to demonstrate that it can deliver the goods that Apple wants. Right now, 10nm chips from Santa Clara are expected in mid-2017, with 7nm not anticipated until 2019 or 2020. 2018 is probably the earliest we’d see chips from an Intel / Apple deal and it wouldn’t surprise us to see that date slip.