Micron manufactures DRAM and NAND flash memory chips. The chips are used in a variety of end markets, including computers, servers, mobile devices, tablets, and increasingly in Internet of Things devices (e.g., connected cars, industrial applications). Micron is headquartered in Idaho and has manufacturing plants in the U.S. and across Asia (China, Japan, Malaysia, Singapore and Taiwan).
Why Does Opportunity Exist?
The memory chip business was historically characterized by high capital intensity and intense competition. Companies spent large sums to build the fabs necessary to manufacture DRAM and NAND chips, and they then needed to operate at max capacity to spread the fixed costs over as many units as possible. Despite robust demand growth, oversupply was a constant problem in a market with as many as 60 manufacturers, causing returns on capital to be minimal if they existed at all.
Today, the memory industry looks much different. The space has consolidated down to a handful of players in DRAM and NAND, of which Micron is likely the smallest remaining survivor. The oligopoly structure provides hope for more supply and pricing discipline, allowing the remaining players to make a decent return on capital in what is still a high volume growth market on the demand side.
Despite the industry tailwinds, Micron's equity has traded off over 70% since the end of 2014. The supposed benefits of the oligopoly have yet to materialize and prices continue to fall double digits annually. Micron is worst positioned as the smallest, highest cost player left. The company is trying to catch up on the technology front to lower its costs, but fears it could be competed away by much larger players like Samsung weigh on the valuation.
Micron is a hairy situation that represents an attractive investment opportunity as several events play out over the next few years.
First, Micron's current earnings are not representative of the business we are buying going forward for two key reasons:
(1)The company has been upgrading its DRAM manufacturing facilities over the past year to use 20 nm nodes, which will bring it in line with the capabilities of Samsung and SK Hynix. The changeover will increase capacity and lower Micron's cost per bit once complete, but in the meantime it has caused a ~10% decline in volumes in a segment that has historically grown volumes 50%+ per year. The updates should be completed in the latter half of FY16, setting Micron up for a significant volume and margin rebound in FY17.
(2)Micron is in the process of acquiring the 67% equity stake in Inotera it does not currently own. Inotera supplies ~35% of Micron's DRAM capacity, and the acquisition will allow Micron to capture the full margin on those chips. The combination was expected to add ~$1.4 bn EBITDA to Micron's consolidated financials, though that has fallen to <$1 bn as Inotera also upgrades its facilities to 20nm. Still, the combined business would generate ~$4 bn PF EBITDA in FY16 despite the depressed DRAM volumes.
The near term completion of the 20nm upgrades and consolidation of Inotera will provide a meaningdul tailwind to Micron's earnings over the next 12-18 months. However, beyond that the key question remains about industry structure: will the oligopoly structure finally result in rational supply and pricing that allows manufacturers to earn a decent (or better) return on capital? The bull / bear is as follows:
Bull: DRAM and NAND are oligopoly markets now. All who are left have little incentive to flood the market with supply to win share at the expense of margins and profit dollars. Samsung and SK Hynix are already messaging lowered DRAM CapEx going forward. Rational supply side actions will coincide with increasing demand for memory, resulting in revenue and profit growth.
Bear: At the end of the day, a bit is a bit, and commodity industries don't generate above average returns for long. Even if the remaining players can play nice for a while, someone will eventually cheat or new players will enter. The latter seems especially threatening given the importance of memory and semiconductors to a country like China, who is eager to enter the market and may not have return on capital as its number one priority. In additon, Micron is especially poorly positioned as the high cost producer in a commodity industry. They're capital investments may catch them up for now, but 18 nm is already launching and they'll always be chasing the next big thing.
While a case can be made for either side, we tend to think the bear case is more realistic. Achieving solid returns in Micron's existing business over the medium-to-long term is an uphill battle, and one should expect a conservative valuation even as the near-term catalysts mentioned above play out. That math results in an ok return, and there are two longer-term upsides that make us excited:
(1)3D Xpoint: Micron and Intel's new memory technology could be a game changer. It seems to be a super-charged version of existing NAND with potential beyond that. Given the ever increasing demand for more memory and better performance, you could see this 3D Xpoint being a large ($5-10 bn+) market opportunity that Micron takes a disprortionate share of as the first mover. The technology is still in ramp up phase and will likely not make a meaningful contribution to the P&L until FY18/19.
(2)M&A value: As one of the few remaining players and a pure play, Micron has real value to anyone who wants to enter the market. The company had a $23 / share offer (>2x today's price) in mid-2015 from a Chinese company. Regulatory concerns would be a hurdle, but even if that is so a potential JV in China could create a lot of value for existing Micron shareholders.
How Could We Be Wrong?
The bear case on the industry may play out in even more negative ways than anticipated, harming Micron's business. In addition, any demand weakness would disproportionately hit Micron as the high cost player given others would be able to supply the now lower market volumes more cost efficiently. Demand growth has slowed down over the past decade as legacy use cases (e.g., PCs) reach maturity, but we believe the risk of a severe slowdown is limited given the huge new growth areas for memory (e.g., Internet of Things).
The company will have a significant debt load ($12 bn+) post-Inotera. If Inotera and / or Micron continue to struggle, the debt load could become a burden and limit the company's flexibility. Existing shareholders may eventually be diluted by existing convertible notes or via equity raises if needed. While dilution is possible, we believe it would have a relatively small impact on the overall share count and returns.
I do not hold a position with the issuer such as employment, directorship, or consultancy. I and/or others I advise hold a material investment in the issuer's securities.
- Completion of 20 nm upgrades
- Consolidation of Inotera
- Further evidence of price/volume stabilization in the DRAM market