Part of the reason for the speculation has revolved around Apple. Intel already makes processors for the Mac; but iPhones and iPads use a proprietary ARM-based processor that has been produced for Apple by Samsung. Given the frosty relationship between Apple and Samsung, the scuttlebutt is that Apple has been looking for alternatives - and that the options might include Intel.
The issue that always comes up when this idea surfaces is that a move into the foundry business would come at the cost of lower gross margins - making microprocessors is an inherently more profitable activity than manufacturing parts for others.
But Credit Suisse analyst John Pitzer took a closer look at this issue, and makes a case in a research note issued this morning that there are reasons to think it would be a smart move for Intel to move into contract manufacturing with Apple as a customer.
"Moore's Law - smaller, faster and cheaper - has been the cornerstone of tech economics for 40 year," he notes. "It is our view that technical, financial and scale issues are driving towards a last-man-standing advantage for Intel relative to Moore's Law and that Intel will leverage that advantage not only in its core business but increasingly as a specialized foundry."
Pitzer notes that Intel Foundry Services (IFS) already has three small, privately held customers - Acrhnoix (which makes FPGAs, field programmable gate arrays), Tabula (programmable logic devices) and Netronome (which makes network processors.) But he says there are a number of other potential customers, including not only Apple but also Cisco, Altera and Xilinx.
"To date IFS has remained mostly a niche effort, but we would argue that as Intel's manufacturing lead continues to widen at 14nm and especially at 450mm, IFS will become a more meaningful driver of top and bottom line performance," he writes. "Investors worry about the impact to corporate profitability from foundry business, especially relative to an Intel and Apple relationship. On the surface, Intel's MPU ASPs of $120 versus the Apple ASP of $20-$25, Intel's corporate gross margin of 60-65% versus TSMC's at 45-50% and Intel's cost/wafer of $10,500-$11,500 versus leading edge wafers at TSMC at $5,000-$5,500 would all suggest that any entry into foundry would be margin dilutive with little bottom line impact."
But he adds: "We disagree."
Pitzer writes that he looked at the impact of a foundry relationship between Intel and Apple using ARM chips "and concludes that the profitability profile is much better than the consensus view."
His view: If Intel were to produce 100% of Apple's demand at 14 nm the company generate an extra $5.8 billion in revenue, 47% gross margins (which is below the 60%-65% of the core business) but with operating margins of 35%, above the core of 28%-32%. He thinks that arrangement would produce an extra 28 cents a share in annual profits, with an ROI of 20%-25%.
INTC is up 11 cents, or 0.5%, to $20.60.
