This last week's Sanford Bernstein conference (registration required) in New York provided investors with a candid look at Micron (NASDAQ:MU) - both the good and the bad. Mark Durcan, Micron CEO, was on hand and his presence definitely added to the drama of the occasion. While much of the hour-long session was a reprise of material from Micron that investors have heard many times before, one response to a question from the audience - deftly followed up on by Mark Newman - disclosed new information that I don't ever remember Micron disclosing in such detail. The topic was NAND, and the question was:
Q: Please discuss the missteps in your NAND business - what steps you've taken to remedy this business and what timeframe we should see the milestones hit.
A: [The good news is that this is] a low capital problem to solve. It's not easy but I have confidence in our ability… we're making changes and over time we'll improve. [It (margin improvement) will not be next quarter but you'll … see an improvement in gross margin over the next two to three quarters and you'll see us move to close the gap in the following couple of years, six quarters - something like that. What have we done? Well we reorganized … we really had one organization trying to do too much. We had a NAND Solutions Group and it was trying to manage the roadmap and technology migration strategy and it was trying to control the strategy for the mobile space as well as the storage space. It was trying to do … more than any one organization could do… [In response to these problems] we reorganized the company. We have separated the non-volatile execution engine form the market-facing memory segments, and we've separated the controller development capability from the non-volatile component capability… We now have a focused controller organization that does controller technology for all the segments and tries to make sure we're … optimizing… We've added a lot of executive talent and senior engineering talent. We have a new guy running our storage segment, Dan Thomas who built the storage business at Dell. We have a new guy running our non-volatile technology business and we have new guy running the controller business, Brian Angel, who we brought over from NVidia…. We are taking steps to make sure we can execute…
- Mark Durcan - CEO, Micron
Durcan then went on to elaborate regarding the gross margin "problem" that Micron is experiencing in NAND and the reasons for it:
TLC (Triple Level Cell) [at Micron] is almost nonexistent. [This compares] to SanDisk's 40% […level which is a] big difference [because] TLC is 20% cheaper to build but is the same value because controllers are so good that the customer doesn't care. Micron missed that move - we did not have the right roadmap. (emphasis by author).
Ok. There you have it, but it's not so bad because the 16nm technology will be riding to the rescue, right? Well, no, not this year at any rate. In response to a follow-up from Newman regarding the 16nm node, Durcan admitted that the rollout today stands at "less than 10%", and will not be completed until the end of 2015.
Let's take a look at the competition. While Micron is in frantic catch-up mode here's what SanDisk (SNDK) was saying about NAND in 2013:
"1Y technology node, die sizes, cost and manufacturing conversion ramp will lead the industry in 2014" Sanjay Mehrotra - CEO SanDisk - 2013 Investor's Conf.
So that was last year. What about this year? ("1Y" is SanDisk's A19 process node which is a 94mm 19nm die - "1Z" is there 15nm node)
- "Our 2D NAND roadmap continues to give us cost leadership with 1Y technology this year and our 1z technology which is starting to ramp in the later part of the year and becoming the workhorse in 2015 timeframe." Sanjay Mehrotra - CEO SanDisk - 2014 Investor's Conf
- "In 2D NAND technology […] we continue to lead the industry - die size, cost, reliability, performance - in every aspect of 2D NAND, we continue to be number 1 in the industry". Dr. Siva Sivaram - SVP, Memory Technology
- "On 2D NAND, [… our 1Y technology] we said we would be going into high volume production in 2013, and we did. So 1Y technology in both X2 and X3, 2 bits and 3 bits per cell, is in high volume production and meeting its intended yields, across all product platforms." Dr. Siva Sivaram - SVP, Memory Technology
- "SanDisk's 1Y (16nm) production ratio will exceed 50% in the second quarter…" TrendFocus DRAMeXchange Survey - May 30, 2014
Summing up, in 2014 SanDisk is not only producing more and smaller (the A19 die is 94nm, while Micron's 20nm process node is 14nm) NAND die, they have also fully implemented and are shipping 3-bit MLC (also called TLC and X3). The overall SanDisk cost advantage? Approaching 30%, with 10% of that coming from the technology node advantage and 20% from the MLC advantage. And that's this year. Next year? The 1Z (15nm) node, which is a significantly smaller die than SanDisk's 16nm 1Y product because it has been shrunk to 15nm in both dimensions, will maintain the cost advantage, whatever happens with Micron's 3D program, until Micron perfects the TLC process. How can we be sure of SanDisk's technology leadership? Good question. I'm quoting from SanDisk's 2014 Investor conference above and their presentation provides greater technical detail to back up their claim. My bottom line on their technical assertions? SanDisk reported 51% gross margins in their NAND business in the last quarter's financial statement. That's credibility enough in my book.
Let's go back to market data. Trendfocus's 2013 survey of the Enterprise SSD market showed Samsung (OTC:SSNLF) with a 30+ share and SanDisk with a 17% share. Micron? Not on the list, falling below STEC who had a .3% share. The recent market momentum has swung even further away from Micron. In this year's Storage Visions's conference, Ryan Smith, Samsung's SSD Marketing Executive, claimed to be shipping 60% share of the "consumer" SSD market for the laptop/notebook/ultrabook segment. What are they shipping into this space? TLC-based SSDs.
SanDisk, whose focus has here-to-fore been almost exclusively on Enterprise SSDs, has announced that they have qualified TLC-based SSDs with the top 10 PC OEMs and will be shipping TLC SSDs into the laptop/notebook market in the second half of this year. Clearly, investments made by Micron's competitors in controller technology are paying off in spades.
So, where does this leave Micron and its investors? Much depends on how long it takes for Micron to fix the problem. They have announced that they will be shipping 16nm TLC in late 2014, so that piece of the solution is on the way to completion. That leaves the controller, which they can build or buy. Either way, it's going to take some time. My guess, and it's just that, is that we are talking about at least a six-month product development cycle followed by six months of integration, certification and marketing work to bring competitive X3 SSDs to market through the OEM channel (which would be at least six months to a year faster than Durcan is implying it will take). Assuming that estimate is close to right, there are several takeaways, presented here in order from near-term to long-term, and from tactical to strategic respectively:
1. Micron's ability to wring profit out of its NAND operations for the balance of this year and well into 2015 is questionable. Here's why:
· Until Micron develops and certifies the X3 controller technology and the TLC process they will not be cost competitive in the consumer SSD market.
· Micron's ability to compete in the Enterprise space will be severely compromised by their lack of good controller technology in this high performance/high durability sector.
· With more profitable SSD channel sales and Enterprise opportunities limited, they have three options, and none of them are great:
i. They can go ahead and ship 2-bit MLC SSDs and cut margins to try to make inroads with the OEMs and in the retail channel.
ii. They can cut production or inventory product.
iii. They can offer product at commodity prices to OEMs who will incorporate the Micron die into their products for sale. This is what they are doing currently with the product that doesn't end up in the WSG and ESG offerings.
2. Intermediate term (2015), assuming that they have executed per Durcan's roadmap, Micron can start to aggressively engage the client OEMs and start selling Micron-branded SSD product. Initially, as the number 5 player (behind Samsung, Toshiba, SanDisk, and Intel), they will need to be much more price competitive to gain business. This will continue to depress NAND business margins through 2015, but builds the base for a 2016 and beyond.
3. In the short and intermediate term, continued success and strong margin performance in the DRAM business is essential. DRAM margins will be covering NAND's meager profits for at least through the end of 2015. Despite this, and depending on the tactics Micron pursues above, the DRAM business may not be able to provide significant top line growth for the company as a whole. As long as DRAM margins stay healthy, Micron will have the financial strength to fund the NAND recovery effort.
4. Longer term, Micron must demonstrate that the company can move from a technology-driven company to a marketing-driven company that really focuses on the solution requirements of customers in order to drive high margin NAND business. This is a well-trodden path that SanDisk and Samsung are already deeply engaged in. This is the ultimate strategic question for Durcan and his executive leadership team. If they are serious about fixing this problem, Micron's current level of SG&A spending will climb by at least 100% from its current 4.3% level and may well triple. (Because Micron has both DRAM and NAND product lines, just matching SanDisk's spend may not be enough.) The good news is that increases in SG&A spend will be more than compensated for, as gross margins climb into the 60% range for both product lines over the course of the last half of the decade.
In conclusion, this has been a difficult article for me to write, as I have been (and I remain) a strong believer in the potential of this company and the industry in which it operates. Perhaps the single most bothersome aspect of the whole situation is that this information, so obviously material to Micron's prospects, took so long to surface. Micron has clearly known about this problem for a long while. Why are the implications for the business just now (and then ONLY because Newman doggedly probed for the answer) coming to light for the investment community? I think we deserve an answer. To be fair, Micron has been saying for some time that they are having problems with "TLC", but the implications of that "problem", to my knowledge, have never previously been articulated. Micron owes its investing community more than offhand and elliptical references to technical issues that it faces when their implications so directly impact Micron's quality of earnings.
Am I making too much of this? I have thought long and hard about that possibility. Upon prayerful reflection, I don't think so. There are several reasons for that.
NAND isn't just A business, it is literally THE business of at least the balance of this decade and most probably all of the next. Micron's own market projections show NAND CAGR outstripping DRAM by 38% to 27% over the next four years (2014 Winter Analysts Conference slide). Other projections, which look longer term to the impact of the Internet of Things ((IoT)) technology wave, show even more bullish projections for NAND growth. The current predictions for IoT storage requirements by 2020 are 35-100 ZB, depending on which forecast you want to believe. (IDC on the low end - IBM on the high end) That's zetabytes, as in 10 to the 21 power. Estimated industry production in 2013? 40 PB, or petabytes, which is 6 decimal orders of magnitude below the zetabyte level. DRAM is a good business but NAND is the ballgame. A miss in this business is not trivial!
In addition to potentially missing out on a very big market opportunity, there is also the possibility that Micron as a company will be damaged long term. The history of the IT business world, and particularly semi-conductors, is replete with examples of companies that have fallen behind and never catch up. Micron could not have two more ferocious competitors for the NAND money than Samsung and SanDisk/Toshiba. They will not be sitting still during the next two years while Micron works its recovery strategy.
Ok, ok, so it's time to build a nice fire and throw all those Micron stock certificates in it, right? No, of course not. As I said above, I remain a strong believer in the future of this industry and, yes, Micron's future in it. Having said that, Micron's multiple is at risk until the company provides compelling evidence that they can deliver great margins in NAND. Folks with substantial long positions would want to consider whether thinning the position might now be wise. Take some profits and be in a position where you can ride this out - playing with house money is rarely a bad thing. Short term, I am concerned for this quarter and Q4. Not that they will be terrible - not at all. The DRAM market is so good that Micron will muddle along nicely, even running on only two of three cylinders.
- Long term, I do believe that a rising tide will lift all boats. Here's the case for Micron, with full knowledge of their weak position in NAND. The industry demand drivers are stronger than ever. Demand continues to strongly outstrip supply. As it builds, even commodity level margins may offer Micron a nice return. The NOR business, even though it is diminishing, will continue to contribute a decent and reliable profit.
- Competitors can make missteps too. Right now there is some skepticism about the short-term success of Samsung's 3D initiative. (Micron and SanDisk have both expressed doubts about its cost-effectiveness.) If it fails, NAND supply could get even tighter and commodity margins could skyrocket.
- Micron's 3D strategy may be a home run that offers the company an overwhelming raw technology cost advantage vis-à-vis competitors. Don't forget, Intel (NASDAQ:INTC) is Micron's partner and with that partnership comes undoubted street cred with 3D technology.
- The Micron DRAM business, already a good one, may become a great one that delivers gross margins pushing past 50% towards 60%. Since two-thirds to three quarters of Micron's business is DRAM, that will overcome a lot of NAND under performance.
So yes, there's still a lot of ballgame yet to play. It's early innings, and NAND industry dynamics offer Micron a soft cushion for this self-inflicted pratfall. As bad as this is, there is no reason to panic, but this is a time to possibly trim positions and assess the Q2 financial results carefully. Micron needs to get its act together in both its NAND business and its investor communications. I will be watching for evidence of strong improvement in both areas. I expect I'll have plenty of company in that regard.
Disclosure: I am long MU. 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.
Additional disclosure: also long SNDK