If you are reading this, I hope you have already invested in Micron (NASDAQ:MU) or SanDisk (NASDAQ:SNDK), or maybe both. Smart move - both stocks have given you a great ride over the last year, with SNDK up 44% in that time period and MU giving you a whopping 216% (YOY thru 2/14/14). Lovely stuff, this, and many if not the great majority of we lucky folks owe that good fortune to Russ Fischer. Russ, as you probably know, made the strategic case for the rise of Micron more than a year ago. "This time it's different," he argued.
But IS it? Really? With Micron sitting well under $10 P/E on a forward looking basis, it's obvious that the bulk of the investing community doesn't believe it. Oh, things might be pretty good now, they seem to be saying, but it can't last. An industry that is notorious for value destruction over the course of its checkered history will inevitably blow up again in an orgy of over-building capacity that will destroy shareholder value.
Those of you who follow MU in SA have certainly seen this case made. And up to just recently Micron's statements haven't helped build the case, either. As commentators such as Electric Phred have noted, Micron shareholder communication here-to-fore has hardly helped support the bullish counterpoint. Arguably, (and Phred made the case very well) MU's shareholder communications have been tentative, confusing, poorly phrased, and maddingly opaque, and that's putting it charitably! In the New Year, however, that has started to change for the better, and in a series of investor and analyst meetings over the course of the last five weeks Micron has crafted a compelling case for their future. In doing that, they make a case for a multiple much richer than the stock currently commands. They have, I believe finally addressed the fundamental question:
Why "this time it really IS different - for the long term!"
This article will elucidate the case Micron execs have made, using their own words wherever possible. It will be one of a series that revisits Russ's seminal argument with an eye to mapping the strategic terrain of the memory industry. In doing so I hope to add some value to the ongoing debate about Micron by moving the discussion from the short term to the long term; from the trees to the forest, so to speak, as we examine the economic motivations that will determine behavior over the next three to five years.
The central elements of this new strategic terrain are:
- The new strategic paradigm enabled by the memory industry oligopoly is durable, and defensible.
- Oligopoly is a common and long-lived business model in the technology industry. The processor business is a perfect example - Intel (NASDAQ:INTC) and AMD. How about logic with Altera (NASDAQ:ALTR) and Xilinx (NASDAQ:XLNX)?
- Micron is consistently and clearly communicating their intent to play its part in enforcing the pricing power of memory oligopoly. (see Micron commentary below)
- Process leadership, not clean room construction, is the key to increased DRAM and NAND chip production POTENTIAL. Ultimately though, the amount of capacity OFFERED to the market is a function of the oligopolistic market model that is now just coming into full view.
- MU's gross margins will grow over time from the current level of 32% in Q1 2014 to levels befitting the swing firm participating in an oligopolistic market - oscillating in a range of 60% to 80%.
- Forward PE will rise along with gross margins into the high teens as Micron's strategy become clear, the firm's willingness to prosecute the oligopoly market model is recognized, and quarterly earnings reflect the unprecedentedly high gross margins that will be obtained.
- For the large OEMs, this new oligopolistic model poses immense challenges that will lead to major changes in the way they do business.
- MU's position makes it a compelling acquisition candidate in the intermediate term, but there is good reason to believe that Micron will remain independent and the MU/INTC JV will remain intact.
- Long term, the threat to the stability of "Memory OPEC" is their ability to manage demand and supply.
The case for the durability of this new strategic paradigm will be made, and in future articles the strategic reaction of the ecosystem will be examined. There are two parts to the argument. Part one is Micron specific and part two relates to any of the other memory suppliers and the OEM ecosystem.
To review: the essence of the case Fischer made was conceptually simple: the memory business, with Micron's acquisition of Elpida, would now become a four-member oligopoly composed of Samsung (OTC:SSNLF), Micron/Intel , SK Hynix (OTC:HXSCF), and Toshiba (TOSFB)/SanDisk. This oligopoly, which we will term the Memory Industry Production Organization (MIPO), would, (as Russ argued) and now do control 95% of memory production. Given this level of industry concentration, each of the four members would be highly incentivized to act (informally, of course!) as a memory cartel. In essence, a new strategic paradigm would govern the memory business. Central to this new paradigm is the notion that industry capacity growth going forward would be gated by two primary factors, one old and one brand new.
The old factor inexorably growing memory supply is the Moore's-law -driven increase in node density governed by technology. Nothing new here, but it is worth noting that Micron is saying that node transition times are stretching well beyond two years. Now for the new factor gating production growth, heretofore unseen in the memory business of old - and that is, irrespective of node density increases, the real driver of growth in the bit supply will be the MANAGED growth in overall end-user demand for memory, especially NAND. This second factor is directly attributable to the new oligopolistic business model, and makes all the difference to the investment thesis for the memory business. The result: no longer should stakeholders in the business fear the rise of exogenous capacity due to new players entering the market; no longer would current industry participants engage in a beggar-thy-neighbor race to build new fabs. Why do that, when maintaining the status quo will lead to reduced investments in CAP/EX and higher memory prices that would lead to an improved bottom line?
This is how Kip Bedard, VP of Investor Relations, put it recently in response to a question from an analyst at the Stifel investors conference held on March 11. Paraphrasing it, an analyst asked under what conditions would Micron consider investing in more NAND capacity?
"What would cause us to add? The one thing we would be watching is that we don't get the cost of the memory going into a system […] where it starts to reduce demand. That's the point for Micron where we need to give customers some relief here. And as Mark [Durcan - CEO] put it at the analysts day that's a long, long, long, long way away from here." [quote edited for readability and emphasis added by the author]
Now, I don't know about you, but for me that statement is amazing. Ripping a page right out of Prince Bandar's playbook, Micron leadership is telling us that the game has truly changed, and that MU, as the swing producer, and MIPO will be managing capacity and demand in the memory business going forward, just like the Saudis managed those two factors in the oil business. So either they are appropriately confident or totally bonkers, right? Which is it? Let's repeat: one year into MU's incredible 200% + run there are a lot of folks out there that do not agree that things have really changed in the business. They are discounting the power of an oligopoly to transform an industry. They think that Bedard's statement is crazy talk. After all, old paradigms die hard. As Mark Twain so famously said,
"We should be careful to get out of an experience only the wisdom that is in it and stop there lest we be like the cat that sits down on a hot stove lid. She will never sit down on a hot stove lid again and that is well but also she will never sit down on a cold one anymore."
After all, if the only thing that is different about today is the level of industry concentration in the memory business, why shouldn't we believe that either:
- One of the Fab 4 will get upset with their position in the status quo industry and fund a large new facility that will provide much more product that end-users can consume, thereby igniting the old race to the bottom with pricing,
- Or, more likely, that an outside player will be unable to resist the temptation of the seemingly easy money and enter the business? How about China? After all, China is notorious for its out-of-control state-owned industry initiatives, several of which are dumping excess product on the world market as we speak. Even Mr. "This time It's Different," our very own Russ Fischer, is paranoid about the possibility of an Intel blitzkrieg of the NAND business.
So what is it about MU's position as a (mostly) merchant memory producer that supports this new paradigm? Let's get a little more granular about Micron specifically. Russ made a strong case that Micron, among the four horseman of the memory cartel, would benefit the most from this new set of circumstances. Why? Four reasons:
- Micron becomes the "swing producer." As the number 2 supplier for both DRAM and NAND (behind Samsung ) and the number 1 merchant supplier, Micron would now have the most "available for sale" capacity in the business, and as the swing producer would be able to adjust its output to enforce the desired pricing environment. In effect, buyers are coming to Micron, not the other way around. Micron is in a stronger negotiating position - a price setter, not a price taker.
- The Intel joint venture. Not only does this JV provide Micron with process leadership in the business (the 16 nm NAND product uses Intel's Hi K/metal gate process and HMC uses Intel's TSV process), it also greatly reduces the CAP/EX overhead for Micron - Intel and Micron split the R&D costs and production costs.
- Industry geo-politics position Micron as the "Switzerland" of the memory business, having no natural enemies who are averse to doing business with Micron and some natural friends who have reasons to greatly prefer doing business with MU. Apple (NASDAQ:AAPL) is among the latter. (Notably it was Apple's mobile Dram contract with Elpida that kept them afloat long enough to avoid liquidation, enabling MU to complete the acquisition.)
- Micron's potentially significant profit leverage from their efforts to move up the value chain. As it stands now, Micron uses about forty percent of its NAND production in its own line of flash memory devices, and has ambitious plans for SSDs going forward. This helps MU increase gross margins in two ways, in that presumably the "value-added product" will come to market at higher gross margins than could be obtained by raw memory, and two, to the extent that Micron takes capacity out of the "merchant" category and into the "captive" category, total memory supplied to the industry is less, forcing the bidding higher for that memory remaining at offer.
So ok, the skeptic might say, Micron is sitting pretty right now, but it won't last. If we're going to answer that question we need to take a hard look at the fundamentals. We need to take a hard look at the oligopoly business model and its ability to endure. One would think after living almost 35 years with OPEC, to mention the most famous instance of this prevalent economic model, that this shouldn't be a hard case to make. In the interests of thoroughness, let's consider point by point how this new arrangement in the memory business fits (or doesn't), so please bear with me while we get granular on this point.
In economics and business law an oligopoly is:
"A formal or informal group of otherwise independent businesses whose concerted goal is to lessen or prevent competition among its participants."
It has the following characteristics:
- Behavior - An oligopoly maximizes profits, not revenue.
- Ability to set price - Oligopolies are price setters rather than price takers.
- Entry and exit - Barriers to entry are high. The most important barriers are economies of scale, patents, access to expensive and complex technology, privileged access to scare resources, and strategic actions by incumbent firms designed to discourage or destroy nascent firms.
- Number of firms - Few - a "handful" of sellers. There are so few firms that the actions of one firm can influence the actions of the other firms.
- Long run profits - Oligopolies can retain long run "abnormally high" profits. High barriers of entry prevent sideline firms from entering market to capture excess profits.
- Product differentiation - (Indifferent from a business model perspective but pertinent in the memory business.) Product may be homogeneous (steel) or differentiated (automobiles).
- Interdependence - The distinctive feature of an oligopoly is interdependence. Oligopolies are typically composed of a few large firms. Each firm is so large that its actions affect market conditions. Therefore the competing firms will be aware of a firm's market actions and will respond appropriately. This means that in contemplating a market action, a firm must take into consideration the possible reactions of all competing firms and the firm's countermoves.
- Non-Price Competition - Oligopolies work hard to avoid competing on price. Loyalty schemes, advertisement, customer service and product differentiation are all examples of non-price competition.
- Perfect Information - this varies depending on which of the four flavors of oligopolistic markets the firm participates in, but in the "non-price competition" model that describes the memory business, information among the participating firms is relatively high.
[ Source Ref: Microeconomics for Today - Irwin Tucker]
A close reading of Micron's statements over the past few months shows conclusively that Micron intends to compete as an oligopolistic firm. Let's take a look at how they support the most important elements of this model with their words.
Behavior: Profit vs. Revenue orientation: DRAM
Micron has emphasized repeatedly that they do not intend to build new fabs anytime in the next few years.
- "Industry Bit Supply Growth is Slowing" - DRAM - From 49% in 2010 to 24% 2014 and NAND - From 74% in 2010 to 37% in 2014 [slide 9, 2014 Winter Analyst Conference].
- [discussing slide 9 above].
"Historically there has been a point to operate on the efficient frontier […] Micron will continue to drive to that efficient operating frontier by deploying advanced technology[…] But the net result of that is a smaller increase per annum in the bits supply […]. In an environment like this [ed. - large increases in per annum demand] it's very difficult for technology to keep pace […] And so suppliers in the marketplace, in order to out run demand, have to go out and add new wafer capacity [ed. - a result which is in conflict with the desire to be at the efficient operating frontier] and that decision is, what is best for my company in terms of gross margin and return on investment and that balance I believe […] will lead to an environment where gross margins can be very good on a sustainable basis." - Mark Durcan, CEO, 2014 Winter Analyst Conference [ed. Bold emphasis added]
Profit vs. Revenue orientation: NAND
- [In response to an analyst question regarding the need to add new fab's for NAND production]
[…] "When we think of about the future and the need for incremental capacity […] to the extent that we think [we need it] we would look to do it on a cost-effective basis but not necessarily with a mind to build out big new fabs. We would look to have clean room space available, as others in the industry are doing, to facilitate transitions and to enable small incremental additions to capacity […]. We've got a low [fixed] cost structure and we've got lots of flexibility in terms of how we deal with capacity to have the right market environment out there." - Mark Durcan, CEO, 2014 Winter Analyst Conference
- [In response to a question regarding 450 mm wafer investment in the next 3 to 5 years]
"Relative to 450, if that ever happened, which is in doubt at this point […] I am not at all convinced that 450 will ever happen […] but to the extent that it does it is a long long way out into the future. […] there is a lot of investment that has to go on and the value is dubious." - Mark Durcan, CEO, 2014 Winter Analyst Conference
- [In response to another question about over all industry capacity]
"…So when we think about our business today [as opposed to a few years ago when we were 10 % of the NAND trade and 13% of DRAM when] our customers were pushing us for more capacity. That's not the conversation we're having with customers today [but rather] how do we use that capacity to enable customer relationships, and this is putting us into a position where we ask how do we get the most out of this [existing] capacity?" Ron Foster, CFO, discussion regarding slide 35, 2014 Winter Analyst Conference
Behavior: Ability to set price
Micron communication regarding this point (which of course is closely correlated to the discussion above) is three fold:
- Most of Micron's product is sold on a contract basis - very little of their business is exposed to the spot market.
- Memory is not a commodity any longer. The evolution of memory from a largely generic PC DRAM commodity to a much more differentiated specialty product composed of DRAM tailored for the server, pc-client, mobile - smartphone and tablet, and networking markets has meant that customers increasingly value suppliers who partner closely to certify product and deliver it in volume and in alignment with new product introduction schedules. Price is only one, and in many cases not the most important criteria is a sourcing decision.
- The customer base has increased dramatically, so no one customer (except, perhaps Apple) has too much leverage sourcing product. (Samsung sources internally.)
- Micron and MIPO are highly motivated to restrain customer demand, when necessary, by maintaining high(er) prices.
- Micron has changed, reflecting all the realities of the points above, in addition to the following factors elucidated by Kip Bedard (VP of Investor Relations) in the Stifel Conference on March 11.
Q. What's different about Micron?
[Ed. - We are a different company facing a different environment.] […] we don't need scale anymore. We have enough scale to adequately amortize the fixed expenses in R&D and SG&A. We don't have to chase scale anymore. The competition has shrunk, […] the end [user] segments have diversified. [So our behavior has changed and that is reflected in, first,] a reluctance to invest in new wafer capacity unless you have a long term view that demand will exceed supply […]. Secondly, much more of our cost basis is variable than it ever has been in the past, so your willingness to run your facility at less than 100% makes more sense today to prevent some sort of major pricing move. Thirdly, we'd be willing to grow inventory because this chase for cheapest bit is […] not with us anymore. [… Yet another factor that is different is our capital allocation strategy, which emphasizes operational flexibility...] Thirdly, […] we've got to be doing more than just be a supplier of chips and bits into the market. We've got to transform ourselves into more of a solutions company and the market wants us to do that. The market is looking for solutions to the problems that they have […].
All of these factors mean that Micron, increasingly, will have the ability to set prices. The shift is not complete, but the factors above will have an inexorable effect on Micron's pricing leverage.
Behavior: Entry and exit
A good argument could be made that the barriers to entry in the memory business today nearly insurmountable for anyone other than Intel. The same commentary supporting Micron's approach to capacity management and pricing is relevant to the issue of barriers to entry in the memory business, which are much higher today than ever before. Three factors should be considered:
Capital Investment and Time to Market
- New fabs are incredibly expensive to build - estimates range from $3-5 B.
- Time to market, given the long lead times necessary to construct the clean room space and populate the chip manufacturing equipment necessary for the node process, is very long, up to 3 years, especially when one includes the time required to bring yields up to best in class levels.
- The technical challenges for new participants contemplating entry to the market are especially daunting considering the entry point for first product shipments are well into the mature first generation stage of 3D product in 2017.
- Most importantly, memory product now is becoming highly differentiated by product segment, and future trends are coupling controller technology in the package in order to meet demanding customer power loading and throughput requirements.
- The Investment Pyrrhic Victory Thesis. Baring a failure to manage market demand and supply by current industry participants by MIPO, any potential exogenous industry participant would face the prospect that, to the extent that their prospective entrance into the market is credible, shareholder value in the new entrant's business would take a massive hit. Gross margins would plunge across the industry, as would return on invested capital, thereby destroying the investment thesis for market entry. (much more on this in the next article)
- Government anti-trust and security interests are likely to significantly complicate any potential acquisition of any member of the FAB 4, thereby adding considerable (if not insurmountable) obstacles in the way of a potential acquirer.
There is much more to be said about the questions raised by this analysis, and I will address those questions in future articles. For now I hope you are convinced, as I am, that:
- Micron sees the business opportunity in the new oligopoly industry structure. Although they do not and would not say so explicitly, I believe they see their role as central to the success and long-term stability of this new structure. They are prepared to act as the "Saudi Arabia" of the new memory OPEC.
- Related to the above, Micron is confident that their JV with Intel will give them process leadership, which means that it is likely grow its share of industry merchant capacity faster than any other MIPO firm. Ultimately, this is the hammer that will discipline other MIPO members to stay the course.
- The durability of MIPO will lead to unprecedented intermediate and long-term increases in shareholder value due to significant increases in gross margins and PE.
- Micron is starting to get the message on the requirement to up their game in shareholder communication. (Thank you, Phred!)
- MU is a stock whose best days are still ahead. Go long and prosper! (Thank you, Russ!)
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.