As one who listened to the Micron (NASDAQ:MU) 2017 Analyst Day who has been following this company for a while, I came away with two principal thoughts. First, and undeniably, there was a lot of good news delivered for Micron's shareholders. The company's leaders made a compelling case in every business line that the future prospects for Micron have never been brighter. <\/p>\n
The second thought followed close on the heels of the first and grows directly out of that good news. Micron can't afford to fund its very bright future. So here on to the message of this brief article: the company's future progress will be hindered if the Board does not act decisively to find a deep-pocketed partner that will enable Micron to take advantage of its emergent strategic technology strength. <\/p>\n
I realize that there is a lot that needs to be unpacked regarding these two items, so much that it simply can't be contained in one single article so consider this piece a strategic summary that will be followed with several more detailed pieces on the key areas that investors need to be aware of going forward. For now, though, here's the big picture. <\/p>\n
Short term (this FY '17), Micron has a license to make lots of money in the memory business. I'm on record predicting $3 for the year, and I now think that is low, perhaps significantly so. Electric Phred's recent piece projecting a possible $4.45 is an excellent analysis of the best-case scenario for this year. It's unlikely we'll see that much but an outcome north of $3 is certainly possible. The question for the stock, however, is how much of that bonanza gets priced into the shares. Here we go again. In the end, in good times and bad, Micron's stock performance is always about the multiple, and the low multiple is always a function of deep and abiding doubts about the ability of the company to overcome the business cycle and show consistent performance. <\/p>\n
I believe the company presented compelling evidence that the long-term growth prospects are excellent. Specifically, the company is capable of taking market share in NAND and new memories over the balance of the decade and well into the next. It will see DRAM share gains this year and next, but it is in NAND and new memories that the share gains will be the most pronounced. For me that's the biggest thing from the day. If the '17 news was good, the FY '18 and '19 news was potentially way better. Here's why. <\/p>\n
Let's look at three of the slides. Exhibit 1 is this slide that rather confusingly expresses Micron's confident assertion of a 25% NAND cost advantage at the 64-layer level, something that DeBoer confirmed in the Q&A in the following exchange: <\/p>\n<\/p>\n
Q.\"Does the 25% die size advantage at 64L translate to cost advantage.\" <\/p>\n
A. \"It absolutely translates to cost.\" [..] We have no concerns with our yield curves on this technology.\" - Scott DeBoer, VP Technology <\/p>\n
The second slide of interest is this one: <\/p>\n<\/p>\n
Here's what really strikes me on this slide with its associated comments: <\/p>\n
The yields on the 32L are equal to planar MLC - probably low-mid 90% range.\"Meaningful\" output on 64L (Gen 2) late this fiscal year - translated - greater than 10% of the 3DN bits will be 64L where the big cost advantage kicks in. Gen 3 (probably a 96L die) is close behind Gen2 with manufacturing start-up this calendar year. QLC. (Translation - 4 bits per cell) This could be the real sleeper of the presentation in terms of its impact on Micron's share gain potential. \"Aligned to market potential\" means that Micron is doing this at the behest of the hyperscalers and will ship it into that market if it meets certain (unspecified) requirements. At the 64L node this technology could take big swathes of HDD in the \"warm\" storage tier. <\/p>\n
Finally, the best slide of all, the slide where Micron claims it has a strategic technology advantage over the long term: <\/p>\n<\/p>\n
This slide was offered alongside a lengthy colloquy by Scott DeBoer that was really quite remarkable for a company as buttoned down and as loathe to hyperbole Micron is (Think of DeBoer as the anti-Rob Crooke). I highly recommend that you listen for yourself (38:10 in the webcast audio), but the gist of it is that bringing a new memory to market is very hard because it takes a long time and it must overtake an old technology that is also advancing. Micron has two new memories that meet that imposing criteria in addition to having the best NAND technology on the market. <\/p>\n
First, 3D XPoint [XP]. According to DeBoer: <\/p>\n
\"This is actually the only commercially ready high density memory technology on the market right now [ ] and there's a lot of interest in this. [But] we're focused on the next two generations and those two generations are going to give substantial cost and performance increases compared to the current [product]. We have high confidence in that roadmap.\" <\/p>\n
This rather innocuous statement needs translating for those of you new to Micron and the industry because it has immense implications. Because of the cloud of mystery that has surrounded XP since its announcement in 2015 there has been great doubt that the product could be cost-effectively scaled. The implication of this claim being that XP would be something of a one-trick pony that would have a relatively short shelf life commercially - if it gained commercial viability at all. The problem you see was that the cross-point arrays were too litho intensive and that wouldn't allow the technology to be scaled. Here's a flavor of the skepticism in the tech community in the form of a slide from Dave Eggleston's presentation at FMS '15. Note the comment at the bottom of the slide: <\/p>\n<\/p>\n
The bottom bullet is pretty much the conventional wisdom in the industry. Here's a Paul Alcorn article from \"Tom's Hardware\" citing the views of Samsung (OTC:SSNLF) and SanDisk (now Western Digital (NYSE:WDC)): <\/p>\n
[] A Samsung whitepaper presented in 2012, [] contends that 3D crosspoint architectures are not scalable in a cost-effective manner over the long term. The lithography tools and the need for EUV for advanced nodes (which IMFT has acknowledged for 3D XPoint) are significant barriers, along with the number of lithography steps. It's noteworthy that SanDisk has also expressed a similar opinion that 3D crosspoint architectures aren't scalable--and it speaks from experience, since it has shipped multi-layer crosspoint architectures in volume in the past. [] IMFT may have the silver bullet to solve the production challenges, but it's apparent that other industry heavyweights are skeptical.\" <\/p>\n
So much for that problem. DeBoer's assertion that he has \"high confidence\" in \"substantial\" XP scaling is enormous news for Micron in the mid-term - think 2019 to 2023. The story that comes together on the slide is the positioning of XP vis--vis ReRAM (resistive RAM), which has generally been assumed to be the natural heir to the mid-range NVM crown. Note DeBoer's positioning - XP is higher performance and less costly than the potential competitive technology. <\/p>\n
Contrast DeBoer's slide with what is still the industry's view of how the future technologies will play out: <\/p>\n<\/p>\n
Source - Western Digital Corporation <\/p>\n
Overlaying DeBoer's chart with the WDC chart, you would find XP positioned to the left and below ReRAM with a performance\/cost oval bisecting the green background. Notice the die cost positioning on the log chart. DeBoer is positioning the future generations of XP significantly less than the ReRAM cost which is shown ranging from $.15 to $.30 per GB (Note: this is NOT where Gen 1 XP is today). <\/p>\n
Here's DeBoer again from the webcast: <\/p>\n
\"The substantial challenge [is to] find an intercept point for that technology where it is meaningful when it gets there. If you look at the commercially viable technologies today and project where they are going to be in five years or so you see a couple of trends here. One is for a new technology to be in this bottom quadrant on the performance versus cost-per-it graph both NAND and 3D XPoint have very significant cost reduction paths over the next few years. So this is very much a moving target for any new memory to intercept [] with a cost profile that makes it significant and enables a market at the same time. So the opportunity there for prospective [replacement] technologies [] is very challenging.\" <\/p>\n
When DeBoer uses the term \"high confidence,\" take it to the bank. The \"very challenging\" comment is as close to smack-talk out of the mild-manned DeBoer as we're likely to hear. Whatever is going on with the XP program right now (I have written about the many mysteries with this program here), DeBoer is giving a full-throated assurance that the best is yet to come. <\/p>\n
And the good news (and bragging) doesn't stop there. Note the \"New Memory\" positioning that overlays DRAM performance at the low end (probably at the \"1Y\" node in 2019) with a cost that is somewhat less expensive. Note also its positioning vis--vis STT-RAM (spin-torque RAM) that has been thought to be the likely high-end DRAM replacement product. As you can see, Micron feels it has a winner that is much less expensive than any STT-RAM offering with higher performance to boot. <\/p>\n
There are two important takeaways from this. First, this is a true DRAM replacement memory, and second, as Micron has made clear previously, this is NOT an IM product. This is Micron proprietary and, if achieved, represents a strategic coup for the company. The bottom line: Micron has the best technology in its cost-leading 3D NAND at the storage end of the market along with two new main memories that will by the early 20s offer customers an unparalleled combination of cost and performance for every use case. <\/p>\n
So what are we to think of all this? Personally, I take DeBoer at his word. Obviously there is a big leap of faith in my belief, because we haven't seen any of these technologies yet. Nevertheless I believe him, and I have a great deal of confidence that we will see the claims made in the slide delivered. I say this because over the last three years, he has delivered - consistently and, more often than not, ahead of schedule. <\/p>\n
One thing DeBoer did not say was when. He did talk about Micron's advantage with 3D NAND and XP being pervasive through a 10-year period. The \"New Memory\" reference is very interesting because it has reappeared after being absent from Micron commentary throughout 2016. Here is DeBoer's Memory timeline from Micron's Summer Analyst Conference in 2014: <\/p>\n<\/p>\n
Originally positioned as \"New Memory B\" for an FY 2017 announcement, I do not believe he would have mentioned it if he did not think he would be announcing sampling in the FY '18 time frame. Assuming that is the case and it has a development and roll out schedule similar to XP, that would imply that Gen 1 would be deliverable in the 2020\/21 time frame. <\/p>\n
Bringing this altogether, the most likely technology scenario going forward looks like this. Take it as you will. <\/p>\n
FY 2017 - XP and Gen 2 64L 3DN initial delivery, \"1X\" (16nm) DRAM initial deliveries, \"1X\" DRAM and Gen 2 3DN > 20% of bit volumes by the end of FY '17 and ramping rapidly. <\/p>\n
FY 2018 - XP market enablers like the \"Gen-Z Consortium\" publish specifications allowing industry-standard XP-DIMMS on the memory bus by mid-year, XP Gen 2 sampling and XP Gen 1 wide deployment with XP DIMM deliveries by year end, Gen 2 64L 3DN bit crossover by year end, announcement of \"New Memory\" [NM] late in FY, \"1X\" DRAM bit crossover, \"1Y\" DRAM sampling. <\/p>\n
FY 2019 - XP Gen 2 bit majority and mass-market deliveries with DIMMs greater than 75% of shipments, Gen 3 (96L) 3DN bit cross-over with Gen 2 3DN by year end, \"1Y\" DRAM > 20% of bit shipments. NM sampling mid-year. <\/p>\n
FY 2020 - XP Gen 3 announced and sampling with XP Gen 2 bit crossover, NM commercial shipments late in the year, 3DN Gen 4 announced. <\/p>\n
Does this look like a challenging pace to you? It does to me, too. By FY 2019, in two years, Micron will likely have four major technologies in production - DRAM, NAND, XP, and NM - with multiple types and generations of DRAM, NAND and XP adding to the complexity. Note I have not addressed the CapEx requirements to exploit this rapid-fire cadence of technology product advancement. Other companies in the industry will be building fabs during this period. Will Micron? <\/p>\n
One other daunting bit of strategic complexity must be added to the list: timing the demise of DRAM. Note the following Intel (NASDAQ:INTC) slide from its recent investor conference: <\/p>\n<\/p>\n
Note the relative size of the DRAM market in 2021 compared to today. For Intel, this is simply market opportunity for XP. For Micron, this is a different and decidedly more poignant issue. Certainly, the good news overwhelms the bad news in this slide. Echoing Steve Jobs, if a company is going to take down the DRAM business, Micron will benefit by being the principal one to do it. Other companies will be bringing DRAM replacement memories to market over the next five years, but Micron is the only one I've seen position a replacement as aggressively as it has. To the extent that it can bring NM to market sooner rather than later, it should be able to harvest profit margins that will cushion DRAM business losses. But this is all speculation at this point. <\/p>\n
So we are left with the question of the way forward. How successfully it can navigate its way through the myriad challenges and complexities of the next four years is certainly open to question. From a technical perspective, I will echo DeBoer's \"high confidence.\" From a business perspective I have my doubts that it can afford to exploit its opportunities. My analysis, which I will detail in the next article, is that the company needs to fund at least $16B in new capacity for the various technologies over the balance of the decade if DeBoer's claims come true. All this with a balance sheet that is seriously in need of repair with $8B of debt and a debt to equity ratio near an all-time \"high\" at nearly 70%. <\/p>\n
Can it make enough money to self-fund its opportunities? The $1.5B of cash flow Maddock is projecting this year will certainly help. Can it achieve equal to better results next year? Most of the analyst community doesn't think so. Many analysts are pointing to the potential of a NAND glut next year with reduced DRAM pricing. I don't agree with the NAND projection for reasons I will detail later, but I can definitely see DRAM pricing receding somewhat. The real issue, though, is that Micron's dilemma remains in good times and only gets worse in bad. <\/p>\n
In closing, as a Micron investor, I couldn't feel better about the technical proofs and claims made in the investor meeting. The short-term business view is very bright. I see $.80 this Q2 and $3+ this year unless externalities overwhelm the secular tailwinds driving the business (#1 on list of possible horribles is a global slowdown driven by a China\/US trade dustup, with the trigger event being a serious run at broad-based import tariffs by the Trump administration). The open question remains funding what could be a very bright future. <\/p>\n
Durcan's departure affords the company an excellent opportunity to use the hiring of the next president (hopefully an outsider) to reimagine the company. However good the balance of 2017 looks, Micron's Board must seize this chance to take a hard look in the mirror. The company has been losing share and is at the bottom of the industry in terms of profitability. Micron has persevered through the PC and Samsung induced trough of 15-16 with a treasure-trove of technology and a boatload of debt. Will the company move aggressively and rapidly to find the resources it needs to fundamentally change its laggard position vis--vis its peers or will it resign itself to remaining in the tenuous position that it has occupied for most of its history? We're about to find out. <\/p>\n
Disclosure: I am\/we are long MU, WDC. <\/p>\n
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. <\/p>\n
<\/p>\n
Originally posted here: <\/p>\n
Technology Rich And Deep In Debt: Can Micron Fund Its Future? - Seeking Alpha<\/a><\/p>\n","protected":false},"excerpt":{"rendered":" As one who listened to the Micron (NASDAQ:MU) 2017 Analyst Day who has been following this company for a while, I came away with two principal thoughts. Continue reading