Micron - No, The Sky Is Actually Not Falling: Highlights Of The Summer Analyst Conference

| About: Micron Technology (MU)


Micron upbeat on demand versus supply in both DRAM and NAND.

Industry output after this year will fall even more than previously projected.

New memory technology is ReRam and it will be additive, not a replacement of NAND and DRAM.

Given the news of the past week with the Samsung (OTC:SSNLF) announcement that the company is repurposing its Hwaseong, South Korea Line-17 logic fab to DRAM production, the anticipation surrounding Micron's (NASDAQ:MU) summer analysts meeting, held in Hong Kong, must have been intense. How would Micron react to the news? How would the company reshape its strategy if, as some commentators on the industry had already concluded, it really isn't different this time?

Mark Durcan, Micron's CEO, addressed the issue in his prepared remarks right up front. Yes, the industry's DRAM output would be up this year, but only marginally up 2016 when the output of Line 17 is expected to come on line. Output will in fact fall more than previously thought in the years 2017 and 2018, due to the technology node transitions that will be forthcoming during that time. Bottom line? As Mark Adams, President, put it when asked about the perception that industry production had "ratcheted up" beyond previous estimates:

"We don't look at it like that. We don't think those numbers have ratcheted up […] We thought the announcement last week [by Samsung] was a validation of what we've been saying all along which is, no one is going to go out and line up a whole new factory. [Ed. - the facility was intended to produce Apple's A8 processor for its iPhone family]. This is an increment into the industry and the fact that it's an in-place facility in a limited environment shows that there's no new massive DRAM capacity coming on line. [The output from that announcement is in our projections] and we're pretty comfortable that the supply demand balance stays healthy. We're very optimistic about that. […] The market will absorb this [new production]. "

Expanding on this topic, both Durcan and Adams emphasized that the big growth in DRAM production occurred through the first half of this year and that they expected production to return to previous levels during the second half. Yes, overall output would rise 30% this year, but it would fall to 23% in 2015, rise slightly to 24% during 2016 (the Line 17 additions finally showing up?) and then would fall dramatically to 17% and 18% during 2017 and 2018. The industry, Durcan opined, is "in a very beneficial and virtuous mode" and that slowing tech node migrations are leaving the market "in line to under supplied". In any case, he said, Micron is "focused on market returns, not on market share". How under supplied? Micron is projecting 25% DRAM demand CAGR against a blended 21% YOY production increase. In other words, the first half of 2014 is as bad as it's going to get. The corresponding percentages for NAND: 38% versus 31%.

Other than the industry DRAM production level discussion, the next generation memory announcements and commentary were perhaps the most interesting and relevant to much of the current controversy on Micron and Intel (NASDAQ:INTC). As close watchers of the industry are aware, Phase Change Memory (PCM) has been advanced as a game changing technology node that will obliterate both DRAM and NAND and in so doing lead to the next great disruption in the industry. Our own Russ Fischer has speculated that PCM could in fact be wielded by Intel to pretty much destroy the current generation of semi-conductor memory producers - perhaps soon.

Micron's take on this theory: not so fast. First off, new memory technology is coming and soon but it won't be PCM, it will be ReRAM. Developed in partnership with Sony (NYSE:SNE), 2015 will see the company's first sample production of resistive RAM or ReRAM with a production ramp in 2016 followed by another iteration in 2017 that will be truly market ready. Will this new ReRAM tech node obliterate DRAM and NAND? In a word - no. Let's hear from Scott DeBoer, Micron's VP of R&D:

"We've got a view that when these new memory technologies come along they are not really displacing NAND […] When we look at new memory in general, […] there is really not anything that is actually competitive with a 3-bit per cell 3-D vertical NAND in terms of cost per bit. We really think it's an additive business possibility. It's not a one plus one equals two kind of thing - it's a two-and-a-half." [Ed - Micron's first 3-D NAND product, being sampled now, is a 32-level TLC implementation]

He goes on to say that the product more likely to be displaced by ReRAM is DRAM, but even there the displacement will be for applications that don't require the full measure of DRAM's speed and low latency and can make do with memory that is much closer to NAND's cost while still being significantly faster.

So, if the new memory technology is not going to threaten Micron, what about the revolutionary new package, the HMC, the Hybrid Memory Cube? Won't the logic manufacturers be the real victors here, especially Intel? Again, no. In response to this question from one of the assembled analysts, here's what Mark Durcan had to say about that:

"Relative to this whole memory processing convergence - I'm a big believer that the future of the electronics industry [… is a …] memory-centric architecture as opposed to a processor-centric architecture and that's because of the diversity of the end-applications that we're serving today, and the diversity of the processing requirements in all those end applications. I think what you're going to see is yes, this is important, but the square inches of silicon in memory continue to grow and the processing requirements relative to the memory requirements of those applications have a very different balance. […] You'll see Micron continue to work on things like Automota Processors and other ways of getting processing into the memory but you'll also see more and more bits of processing that don't require the absolute leading-edge logic process get embedded into the memory as opposed to the other way around." There will be some amount of memory on-board processors […] but the bigger trend is the other way around in the Internet of Things […]."

In conclusion, there is so much more to discuss about Micron's disclosures during this analyst conference, but let's close by adding some perspective to what was said. First, Micron comes across in this presentation as a very confident company. The executive team, from Durcan on down, communicates an air of optimism and excitement. They have made and are making big technology bets and they think that they've bet right and will come up winners. What are those bets?

  1. Their NAND transition strategy. Specifically, the decision to make the 16nm planar node their last planar node before 3-D NAND. They believe that they will have the lowest-cost NAND product on the market through 2015.
  2. Their 3-D NAND approach. They believe that in regards to timing, cost, and performance they will have the best product available on the market at the best price at the right time - 2015 for first volume shipments.
  3. Their DRAM catch-up strategy. They know they are behind Samsung on the 20nm node, but they think that they can overcome Samsung's lead by 2016.
  4. Their next generation technology choice and timing. They think their ReRAM strategy is the right choice and that it will be additive to their NAND and DRAM business.
  5. They are convinced that MIPO, the oligopoly I have termed the Memory Industry Producers Organization, is durable and that there are no existential threats on the horizon from either new entrants or from established logic producers like Intel.

Could they be whistling past the graveyard in regard to the technology threats discussed above? Sure, but as close as Micron is to Intel I've just got to think that there is no way Intel is going to spring any surprises. If Intel really thinks that they can take over the business through their implementation of a "processor-centric" technology architecture, Micron has certainly not received the memo. Since Micron and Intel in fact worked "closely" together (the word that Scott DeBoer used describing the partnership to adapt HMC to Intel's next-gen Xeon server chip), you'd certainly have to think that Intel's capabilities would be well known to Micron, and Durcan wouldn't have said what I have quoted him saying (above). The other thing that strikes me is that much of what Micron talks about in regard to next gen memory technology being ReRAM echoes the same points that have been made by SanDisk's (SNDK) CEO, Sanjay Mehrotra in their 2013 and 2014 analyst conferences.

All things considered, I came away from the call heartened about Micron's intermediate and long term success. Macro factors aside, Micron is positioned to grow revenues and profits and cash for the foreseeable future. Are there some challenges? Certainly, and I will discuss those in the next article. Stay tuned and stay long Micron.

Disclosure: The author is long MU. The author wrote this article themselves, and it expresses their own opinions. The author is not receiving compensation for it (other than from Seeking Alpha). The author has no business relationship with any company whose stock is mentioned in this article. Also long INTC and SNDK