Micron (MU) hosted their winter analyst meeting last Friday, February 7, 2014.
Some of the things that struck me as significant:
Early into the presentation, Micron CEO, Mark Durcan, mentioned, confirming my contention for several months, that Micron is actually the largest non-captive merchant market supplier of memory, period. So, Samsung from a broad customer viewpoint is less important to the industry than the new Micron. Mark also mentioned for the first time that the fact that Samsung competes in the equipment business with many of their memory customers; those customers' preference to deal with other-than-Samsung suppliers is apparent in the market.
Why it is "different this time" is vividly shown in slide 31. In 2004 the memory business served two discernible business segments, PCs and Networking, both of which were commodity, industry standard products…corn, in other words. By 2014, only ten years later, memory served seven highly diversified business segments, many of which have different and stringent requirements that allow for supplier differentiation and, therefore, higher margins, and sharply reduced potential of "bloodletting" among suppliers.
Slide 17. On a 32% gross margin, Micron delivered 22% net margin. Only 12% ($485 million) in op ex. Revenue is likely to increase by 25% while op ex is expected to increase to about $550 million or 13% of sales. Modeled at a $5 billion 40% GM quarter, the company could deliver $1.49 billion in net profit or nearly 30% net. That could be a $1.35/share quarter.
Here's a tidbit; Mark Durcan talked about the storage capacitor being the limiting factor in smaller DRAM nodes:
The challenge that actually happens when you start going at some node, whether you want to call it 1Y or 1Z the challenge with that what's happening is you're getting to the point where you can't fit the number of films that it takes to build the capacitor inside the space that you have available for the capacitor. So, you either have to come up with new films that can be substantially thinner, which is a possibility, or you have a capacitor problem."
So, I think Mark is alluding to High K films for making those capacitors. HK technology is probably learned (transferred) from Intel (INTC) through the Intel/Micron joint venture.
GDDR5 is being brought to the graphics party by Elpida. This will be a "must have" for game consoles and even HPC. Not easy to make and not cheap.
HMC will move the needle in 2015. 18 design wins. HPC was the target application for HPC, but interest from the networking segment has been a pleasant surprise. This is TSV (Through Silicon Via) technology, an important technology for the future.
16nm NAND is looking good from a yield ramp. Slide 45: "Exceptionally cost effective 16 nm will be majority of Micron's output by 2H 2014." This is extremely good news because it could reduce the chip size of a 128Gb chip to 120 sq. mm. This shrink could make a 256Gb chip possible at 200 sq. mm. or a little more. The success and economics of the 16nm NAND process is actually pushing out the economic viability of 3D for all suppliers, including Micron. 16nm is already qualified for use in SSDs.
40% of Micron NAND is finding its way into systems or subsystems, primarily solid state drives.
From another source:
"Next Chuck Dennison. Senior Director Process Integration, from Micron, provided an overview of where the company is today in terms of its own NAND memory technology roadmap.
Our current generation is 16nm NAND that is now in production and we're showing that it is getting to be a very competitive and very cost effective technology," according to Dennison. Micron's new 16nm NAND process provides the greatest number of bits per sq. mm at the lowest cost of any multilayer cell (MLC) device. Eight of these die can hold 128GB of data. the 16nm storage technology will be released on next generation solid state drives (SSDs) during 2014. SSDs consist of interconnected flash memory chips as opposed to platters with a magnetic coating used in conventional hard disk drives (HDDs).
"Our next node is a 256Gb class of the NAND memory. Technically it could be extended before taking the full step to 3D NAND."
Today NAND is the lowest cost-per-bit memory technology and this continued cost-per-bit reduction is really driving the whole of the NAND industry, according to Dennison. It is why NAND replaced DRAM in terms of total dollars and has continued to proliferate across various applications, and is responsible for continued innovation in portable consumer electronics, such as tablets, where so much functionality enabling photography, video recording, storage of an entire music library, and so on, can be packed into one device.
Outlining Micron's technology scaling path, Dennison explained: "We went to high-K/metal gate to 20nm and we used the same technology to extend us to 16nm. From there, the company is moving to a vertical channel 3D NAND for a 256Gb class."
Of course, Elpida gave Micron a front row seat in the mobile DRAM market, and continues to deliver solutions in that arena. Just before Christmas, Micron delivered the industry's first mobile LPDDR4 samples to the chipset (application processor) manufacturers and the mobile device OEMs. These parts will deliver much higher bandwidth while sipping energy.
Packaging is particularly important in the mobile business, since most smartphones combine the application processor and DRAM in a stacked multichip package.
With the 16nm NAND and early mLPDDR4 DRAM, Micron becomes a one stop shopping center for mobile memory and they don't compete with their customers in the device market.
I find it interesting that there was no mention of Intel during the analyst meeting, and I suppose that is on purpose. These two companies have cooperated for many years. I am certain that the transaction that sold Intel assets in the IMFT joint venture back to Micron for a fraction of their replacement value was a "favor among friends" to bolster the Micron balance sheet and thus enabling the Elpida acquisition.
The HMC was a joint development of Intel and Micron.
20nm and the follow on 16nm HKMG NAND processes were a cooperative effort involving Intel.
Just guessing here, but I suspect that the surprising low cap ex of Micron is to some extent due to Intel making slightly used fab equipment available at "kin-folk" prices. Intel has a full corporate subsidiary whose mission is to sell slightly used fab equipment.
The acquisition of Elpida was a financial "steal" and it was also a technology "steal" and a perfect fit for the weak points of the "old" Micron. Elpida brings Micron into the mobile world. The mobile device business could be slowing, but the memory content per device, both DRAM and NAND, is increasing rapidly.
Micron is suddenly the memory technology leader with the best planar NAND, a plan for 3D NAND, and wide variety of DRAM both in compute and in mobile. Micron is multi-year growth story from here.
I think that there is enough doubt about Micron, memory in general, and the concept of "it is different this time" that will make Micron a great investment or trading play for at least a couple of years into the future.