Intel (NASDAQ:INTC) is rapidly developing into a potential one-stop shopping center for Apple (AAPL) components. Until very recently, the company couldn't supply leading-edge (LTE) baseband and transceiver chips; now it can.
Until the announcement of cost-disruptive 3D NAND technology as recently as November 20, 2014, Intel had no position at Apple on NAND components and NAND-based solid-state drives.
During the past year, all this has changed as we look at each component category.
X86
Since June of 2005, Apple has been using the Intel X86 family of CPU chips in Macs and Macbooks.
From time to time there are rumors of Apple making its own ARM-based CPU chips, but nothing ever comes of it. In the Steve Jobs biography, the Power chip to X86 conversion was described as a very high-risk six-year project involving crack teams from both companies. It was done in total secrecy.
The Mac business is growing for Apple, and is a very important part of the "ecosystem"; to change the brain of the Mac would be insanity.
X86 in Macs is about a $4 billion business to Intel.
Foundry
At this point, Intel is the only semiconductor company that we know, for sure, to be shipping high volumes of 14nm products. Some think Apple will stay on 20nm planar for the A9 chip, since there is no sign of trustworthy production of 14 nm from any foundry.
This is supported by recent softness at the equipment companies for the very machines necessary to produce finfets.
This could put Apple in a large bind. Imagine if the foundries really can't get to finfets at 14nm and Intel is shipping compelling mobile solutions, at 14nm and even 10nm, while Apple is stuck with the struggling foundries at two nodes behind.
The safe decision for Apple would be to arrange to have Intel do the foundry work for Apple. Brian Krzanich has now said that the Intel foundry business is open to all customers, which might not have been true under the previous regime.
"A" chip foundry business, at $37 per unit, could amount to $11 billion.
Solid-State Drives
Intel is making SSDs. Soon to be made with "cost-disruptive" 3D NAND. Apparently, SanDisk (SNDK) has lost the Apple SSD business over quality issues.
SSDs are a $2-$3 billion business at Intel, and growing.
Mobile NAND
The iPhone now starts at 16GB of storage, then jumps to 64 and 128 GB. There is no 32GB. The 64GB model seems to be the big seller. With 128Gb chips, that would be four of those chips stacked in a multi-chip package. That's about a $36 part, with probable yield problems. With the Intel 256Gb chip, this becomes a two-chip MCP with better yield and, since the chip has a breakthrough cost, a lower price.
Apple probably buys $6-8 billion worth of discrete NAND annually, and Intel (and Micron (MU)) apparently are setting up for a significant cost advantage on 3D NAND.
LTE
Since the Apple A series of application processors does not yet have an integrated LTE function, it must be bought as a discrete device. These parts have been supplied by Qualcomm, but could just as well be supplied by Intel.
At ~$30 per chip, this is $6 billion for the iPhone alone plus the iPad usage.
Obviously, any imminent Intel-Apple business relationship is total speculation, but with the ability to supply over $100 worth of critical components on each iPhone built, Tim Cook must be thinking about his supply chain.
Apple's secrecy policy with regard to suppliers is stuff of legend; a supplier CFO, in moment of openness, once told me that their internal policy was that the very mention of the word Apple internally was grounds for termination. When they spoke of Apple, it was referred to as "the Fruit Vendor."
Given the complete opaque nature of the supplier-customer information, Apple and Intel could be on the verge of marriage, and we would never know until it was underway. No forecast, no guidance, just a lot of investor and analyst head-slapping when it happened.
Of the five products and processes mentioned above, Intel has a leadership position in four of them. Leadership in X86 is obvious, leadership in SSDs and NAND through the "cost-disruptive" 3D NAND technology, and the ever-present leadership in manufacturing node on foundry logic. When the LTE manufacturing is brought in-house at Intel at 14 or 10 nm, the communication leadership will be there as well.
Apple can keep betting the company on the often questionable ability of foundries to stay only one node behind Intel, and it can keep funding the competition by buying product from Samsung (OTCPK:SSNLF).
Apple could watch these leading Intel technologies go to its Chinese competitors, or it could get on the Intel bandwagon and streamline its supply chain, without doubt and without competitors as suppliers... and maybe save some money in the process.
What would you do?
