Intel (NASDAQ:INTC) makes Solid State Drives (SSD). Judging from comments made during the recent Intel investor meeting, Intel LIKES the SSD business. According to Diane Bryant, Senior Vice President of the Data Center Group, Intel is the number one supplier of SSDs to the data center business. She also mentioned interest and work in other non-volatile memory technologies. Much of Diane's presentation involved storage in the data center business. Storage is simply another word for memory, usually of the non-volatile persuasion.
In his presentation, CFO, Stacy Smith referred to NAND (slide 32) as, "dear to my heart". He mentioned that the "Non Volatile Memory Solutions Group" was profitable and expected to double in 2014.
So, NAND memory and SSDs are alive and well at Intel. Let me repeat that in a different way, Intel IS IN the memory business.
Currently, Intel receives the NAND memory chips for the SSDs from a joint venture (IMFT) with Micron Technology (NASDAQ:MU). Intel owns 49% of the joint venture and either party can buy the other out entirely by exercising an option to do so. The Lehi UT plant is the IMFT fab. Let's assume the fab is capable of 700,000 wafers of NAND per year. Intel would be entitled to about half of those wafers at a price that, "approximates cost", according to every mention of the IMFT JV. Cost on those 350,000 wafers (at $1500 each) would be about $525 million; made into SSDs at 60% gross margin, the wafers would turn into about $1.3 billion worth of SSDs. The level of Intel SSD business is now about $1 billion and, according to Stacy Smith is expected to double in the next 12 months. At $1.3 billion Intel runs out of supply of those great 20nm and 16nm High K metal gate NAND chips.
As I mentioned in an article published yesterday, there are about 550 million Hard Disk Drives (HDD) sold each year, half of which would really like to become SSDs at about an average density of 192GB. To summarize the arithmetic in that article, the HDD to SSD conversion effort is about five million NAND wafers short of doing the job. That is if you only consider the current and traditional manufacturers of NAND memory.
My other continuing theme in SA articles has been that Intel has a gargantuan amount of real and potential fab capacity that apparently will never be used. Stacy Smith used up a large part of his presentation trying to persuade us that Intel is running at 80% of capacity.
The world's largest and second largest semiconductor fabs are located in Oregon and Arizona and both are owned by Intel. Neither of these enormous fabs have produced even one wafer yet. So, obviously these fabs could not have been included in the capacity utilization calculation cited by Mr. Smith. If they were, the Intel utilization rate would be somewhere in the 40% area and investors hearing that would quickly take the stock price down by half.
I have a theory about what might be going on here.
Consider that Intel likes the SSD business AND Intel will very soon run out of the NAND chips it needs to continue to expand the SSD business AND Intel has what might be as much as three million wafers of excess capacity sitting in Oregon AND Arizona. Those "ands" in caps are in the Boolean sense. The only reasonable output of the above three input Boolean expression is the output that… drum roll, please… Intel will use those huge semiconductor fabs to build not just NAND memory, but finished SSDs to begin to satisfy the soon to be burgeoning demand for the advantages that SSDs bring to the computing continuum. This would include near instant boot time, extreme speed performance, low power, ruggedness, etc.
This would certainly bring the growth that Intel shareholders and Intel analysts have been clamoring for.
The other part of my theory is a little more difficult to justify and, therefore I offer it as an "option".
I think Intel will ultimately, and maybe very soon, purchase Micron in its entirety. The counter to that is, "Why go through all that when Intel can just make all the SSDs it needs by itself?"
The answers are:
1. While Intel can configure those new plants to make SSD drives very efficiently and at very low cost and quite high margins, they still need DRAM and non-SSD NAND to go with the Application Processors they intend to swarm the tablet and smartphone markets with. There is at least as much revenue and profit in the memory that goes with an application processor (AP) as there is in the AP alone. And the memory is almost always mounted with the AP in a multi-chip package (MCP) or a package on package (PoP) module. The memory revenue might even be necessary to fully justify going after the mobile business at all. Intel having access to APs with memory would certainly give Qualcomm (NASDAQ:QCOM) some heartburn and give Samsung a run for their money. Intel without memory means they must buy it from a probable competitor.
2. Micron would add the scale and top line growth that Intel desperately needs, and any purchase at under $40 per Micron share would be immediately accretive to Intel earnings. If done with Intel shares, the acquisition would be cash free. Remember, Interest rates will not always be near zero.
3. Here's the more subtle reason: While Intel is the largest semiconductor company in the world by sales, they are about number four in terms of wafers processed . Samsung (OTC:SSNLF) is about twice the size of Intel in terms of wafers out. The new Micron would produce more wafers. Hynix (OTC:HXSCL) and Taiwan Semiconductor Manufacturing Company (NYSE:TSM) also process more wafers than Intel. So what, you say. Just the fact of processing wafers in huge volumes leads to the disciplines that make process yields better. The semiconductor industry has always used memory as the process driver and verifier. Without the highest volumes, the companies processing larger volumes of wafers will eventually overcome the Intel manufacturing advantage. And, of course, there is the scale factor involved in the cost of raw wafers, chemicals, and the manufacturing equipment that is the lifeblood of advanced semiconductor processes. Intel doesn't want those advantages to go to Samsung by default.
Intel has the opportunity to develop and take a $35 to 50 billion SSD business. They are the only one with enough empty capacity to get the job started. Intel also has the opportunity to gain another $16-20 billion of scale by buying Micron at an earnings accretive price.
Combined with the base X86 business, Intel could go on a growth spurt to $100 billion and beyond. At that scale Intel would earn $25 billion, spread over about 6.5 billion shares or $3.85 per share. These stars are lining up and they won't stay in alignment for long. If Intel doesn't move on these opportunities within six months they will have a tough time from here on.
My play will, as always, be with options. If a deal isn't obvious by June, it will never happen. So, buy the stock or June calls and be gone in June if nothing happens.
Disclosure: I am long MU, INTC. I wrote this article myself, and it expresses my own opinions. I am not receiving compensation for it (other than from Seeking Alpha). I have no business relationship with any company whose stock is mentioned in this article.