Last month, the New York Times shocked the semiconductor industry with an article detailing an elaborate heist of corporate secrets from Micron Technology's Taiwan office. The article, and later coverage in the Wall Street Journal uncovered systematic corporate espionage by Chinese state-backed entities seeking to jumpstart China's semiconductor industry. In the weeks since, Idaho-based Micron has become a flash point in the budding trade war between China and the U.S.
Eric Gregg, the founder and principal of Four Tree Island Advisory, submitted a comprehensive long report to SumZero on Micron. SumZero sat down with Gregg to discuss Micron, the semiconductor industry and the implications of a trade war between the U.S. and China.
Luke Schiefelbein, SumZero: What about Micron initially caught your eye as a value investor? What sparked your interest in the name?
Eric Gregg, Four Tree Island Advisory: My initial interest was perked due to my work in the Wafer Fabrication Equipment (WFE) sector. I've become a strong believer in a) the long-term secular growth dynamics gripping the WFE sector b) that there is a sizable valuation disconnect between top companies and their underlying fundamentals, and c) that the outlook for memory is fundamentally better than what the market is buying into (which provides strong underpinnings for the top etch and deposition WFE companies). Given Micron's historically volatile operating and stock price history and at times seemingly over-leveraged capital structure, it took some work to become not only comfortable, but enthusiastic about the potential for Micron's stock. I've come to believe that Micron is fundamentally a different company now than it was just three years ago due to the robustness of its end markets, its tri-opoly industry structure (at least in DRAM), its significant technological gains vs. the competition and the increased capital and technical complexity that are driving meaningful barriers to competition. Throw on top of that that the stock is the cheapest stock in the S&P 500 based on forward earnings and it reads as a highly compelling investment opportunity.
Schiefelbein: Your initial report (12/18/17) covered the various outcomes for Micron given the potential of material disruption to DRAM/NAND production in Korea. Is there any likelihood of disruption of the Korean DRAM/NAND supply chain similar to Thailand’s 2011/2012 flooding?
Gregg: That was one of the three potential levers (the other two in the report were the potential for something strategic to happen with Intel (the impetus being some buyout provisions that kick-in in January of 2019 around the Intel Micron Flash Technologies JV) and for company and industry fundamentals to continue to improve and the stock re-rate). While the likelihood of a disruption in supply due to conflict on the Korean peninsula appears more remote now than it did back in December of 2017, a) it is still a possibility (the WSJ just this week reported that North Korea continues to expand a key missile manufacturing facility and the message Secretary Pompeo conveyed during this week’s visit to North Korea was not well received) and b) we now have a different conflict (a trade conflict with China) that is having a more meaningful impact on Micron specifically and the semiconductor industry in general.
Schiefelbein: What catalysts should investors pay attention to as your thesis plays out?
1) The biggest overhang on the stock is the China trade conflict situation. The main China trade related concern at the moment is there is a commercial dispute that China appears to be using as a means to force technology transfer and inflict pain on Micron as retribution for the US’s treatment of ZTE. If/when this issue is resolved, that should be a major catalyst for the stock.
2) Continued strong operating performance
3) DRAM and NAND pricing per bit that doesn’t drop more than costs per bit
4) Continued indications of a robust global economy (Micron and the semis are considered to be a pro-cyclical investment play despite there being strong evidence of some important secular growth and individual company repositioning themes at play for Micron).
Schiefelbein: Why is memory pricing so volatile and cyclical? Do you expect DRAM/NAND memory to continue towards commodification?
Gregg: There are two different markets: DRAM and NAND. DRAM is showing strong signs of being a much more rational and disciplined market. Growth forecasts are for 20%+ annual bit growth for at least the medium term. Shrinks in DRAM are becoming notably more complicated and expensive. For all practical purposes there are only three sizable DRAM manufacturers and there are serious questions about whether new players (mainland Chinese) will be able to successfully compete in leading edge DRAM. On the NAND front, medium term forecasts suggest a 40+% bit growth rate. With higher anticipated growth, more ways to achieve that growth (moving from TLC (tri level cell) to QLC (quad level cell), for instance, provides a 33% increase in bits per chip – moving from 64 layers to 96 layers provides 50% more bits per chip) and more major competitors in the space (currently six), there is decidedly more risk of continued commodification of 3D NAND. As the dynamics in DRAM appear to be transcending to much more healthy and disciplined industry behavior (and DRAM is typically 70%+ of memory revenues for the major players and over 80% of operating profit), I believe the industry structure is changing significantly for the top three memory semi manufacturers (Samsung, SK Hynix and Micron).
Schiefelbein: How could the recent protectionist American shifts in trade policy impact Micron (with ~50% of sales made to China)? What is your best / worse case?
Gregg: This is, without question, the billion dollar question and topic of the moment for Micron.
There are three, principal areas of risk for Micron:
- The US limits the sale of Micron’s chips to China (I discuss this in my update and think this risk is low)
- China fines Micron and the other DRAM memory semis for charging too much for DRAM (I discuss this in my update as well, hard to estimate the likelihood of this one maybe 50%/50%).
- China enforces IP to a) protect the IP it stole from Micron by using its enforcement to compel Micron to license its IP to China and b) harm Micron as retribution for the US’s treatment of ZTE (this appears to have happened).
According to a press release from UMC on Tuesday, a Chinese regional court ruled in UMC’s favor and a preliminary injunction has been issued enjoining Micron from selling 26 of its products in China. Micron Thursday issued a news release acknowledging receipt of the ruling, but also highlighted that the 26 products impacted by the ruling only account for ~1% of Micron’s revenues.
Before going into the best and worst case scenarios, just a little background. UMC was hired by Fujian Jinhua Integrated Circuit (which is about finished building a $6 billion semiconductor plant according to the NYTimes) to provide the technological specifications it needed to make memory semiconductors. UMC has no recent history of producing leading edge node memory semiconductors. According to Taiwanese police reports and a criminal indictment of UMC in Taiwan, the Company stole the IP (at the direction or at least consent of Fujian Jinhua IC) for its memory chips from Micron and is now, according to Micron, in turn suing Micron for IP infringement as a “proactive defense” due to Micron’s efforts to bring UMC and Fujian Jinhua IC to justice for its misappropriation of IP.
What makes this situation especially absurd is that:
1) According to Micron they have been found guilty in Chinese court without ever having been allowed to present a defense (due process, at least in this instance, is apparently an alien concept in China).
2) Micron asserts that the patents being enforced by UMC are not used in Micron’s DRAM and NAND technology or products.
3) Chinese courts don’t recognize “conflicts of interest” as a concept. With Fujian Jinhua Integrated Circuit owned by industrial companies that are owned by the regional government whose courts heard the case (and the Chinese courts exhibiting a poor record of siding against de facto state owned enterprises), the set-up for Micron to have an initial positive outcome was, in retrospect, an uphill battle.
4) It is clear based on Micron’s press release that it hasn’t been determined by the Chinese authorities yet whether these patents that the Fujian court has ruled as having been infringed are even valid patents.
The fact that the ruling hit the market on the day before the 4th of July has been construed by some as a clear swipe directed at the Trump administration and both a) retribution for what has transpired to date with ZTE and b) an indication of the types of measures China will take if this trade conflict continues to ensue.
What also needs to be appreciated is what this case probably means to China (or at least the Fujian regional government). This case was filed in China a month after Micron filed its complaint against Fujian Jinhua Integrated Circuit and UMC in the Northern District of California which could have global ramifications for UMC and Fujian Jinhua if the case is resolved in Micron’s favor. China’s court system is typically faster than the US’s, but this case has moved with remarkable alacrity. Given the background that UMC and Fujian Jinhua clearly (based on Taiwanese police investigations, their criminal indictment, and Micron assertions) compelled new hires to steal from Micron and then purposefully destroyed and hid evidence from authorities on a large scale, China (or better said, the Fujian regional government) has effectively attempted to “cleanse” their illicit behavior through their state controlled court system (justice is not blind in China). With it likely to take at least six to 18 more months for the US judicial system to conclude its process, China/UMC/Jinhua have given themselves a) leverage with Micron b) China/UMC/Fujian Jinhua can purport to be free and clear of wrongdoing because their court system has ruled so (and not suffer the embarrassment of being convicted by the US court system before being partially exonerated by the Chinese judicial system) and c) China has sent a clear message to other Chinese companies that this type of activity (stealing trade secrets and then suing foreign companies that generated the IP for infringement) is not only condoned but will be protected by the State. The most likely desire of all the Chinese parties involved is for Micron to a) drop its suit in the Northern District of California and b) an agreement by Micron to license its tech (that which was stolen and ideally other core tech that Micron has developed) to its emerging Chinese competitors.
There are a number of ways this situation could unfold:
- Micron can appeal and actually defend itself and any injunction could presumably be suspended until the outcome of the appeals process.
- There’s a reasonable argument that could be made that the Fujian regional government/court ran amok in this situation and that the greater Chinese government could reverse this dust up. It appears that this Fujian regional court is playing fast and loose and a) not waiting to hear whether or not these patents are indeed valid before ruling in favor of its own interests (typically validity is determined before infringement) and b) not allowing the defense to present its case, again, to the benefit of its own interests.
- If Micron succeeds in its case in the Northern District of California against Fujian Jinhua and UMC, that could in turn lead to significant leverage for Micron over those two entities. It is very possible that the Chinese judicial actions may help accelerate the case in California’s Northern District. Fujian Jinhua IC is only expected to be up and running with memory products come 2019. Injunctions could be placed on any products sold into the US market (or potentially even globally) that incorporate a Fujian Jinhua IC semiconductor and penalties could be imposed. Also, UMC trades in the US and various penalties could be imposed on them.
- If Taiwanese authorities succeed in their criminal prosecution of UMC, that could also further improve Micron’s likelihood of prevailing.
- It is clear based on Micron’s press release that it has yet to be determined whether or not the patents in question are valid. While the regional court can determine infringement and prescribe injunctions and other penalties, only the State Intellectual Property Office in China can determine validity. Typically that takes a year or two and this case has only been in the courts for less than six months. According to Micron, the patents being enforced are invalid “because they are directed to technologies that were previously developed and patented in other countries by other technology companies.” China is establishing very bad precedent if Micron’s contentions are accurate and China upholds the patents regardless.
- If this is simply China playing dirty (fixing the legal outcome despite the merits) to a) retaliate against the US for the ZTE situation and b) to penalize Micron for not acquiescing to as much tech transfer as Samsung and SK Hynix (Micron’s investment in manufacturing capacity in China is much more limited than Samsung or SK Hynix), that in many ways plays right into the Trump administration’s hands and provides further evidence of bad Chinese trade behavior. This situations could a) lead to a further ratcheting up of trade tensions or b) help force a better broad trade agreement that also involves a more positive outcome for Micron.
- Given that this ruling only impacts ~1% of Micron’s business, the initial impact on Micron is negligible. But if this type of state and judicial sanctioned misappropriation of IP with subsequent IP enforcement on foreign companies is allowed to stand, it could present a very slippery slope.
Having presented some of the issues on the table:
Best Case Scenario – US and China come to some kind of agreement Micron goes back to operating per normal (with Micron possibly having to agree to build more capacity in China). It is impossible to know how soon this could come to pass, but I believe that both sides (US & China) want to come to some kind of resolution sooner than later. China’s broader stock markets are down ~20% YTD. China’s economy is slowing while its debt markets are showing signs of strain. China has had to inject significant emergency liquidity into its financial system to calm fears brought about by this trade row. The Trump administration’s approach to improving trade balances are facing blowback from top congressmen from both parties, the US’s allies, from some major US employers (like GM) and from some of his core voting block. This all could come to some kind or reasonable resolution sometime over the next few weeks/months and sufficiently in advance of the mid-term elections.
If some material headway were achieved between China and the US to ratchet down trade tensions, that would be highly positive for Micron.
Worst case scenario – actions, not just rhetoric, on the trade front escalate in a meaningfully negative way causing the next global recession. Micron obtains a hard ban on selling any products to China (and the Company has to sell its supply previously oriented to China elsewhere at lesser prices because of a weakening global economy). Investment in the cloud, big data, AI, IOT, increasingly smart driving technology, higher memory content in mobile devices all decelerate significantly. Long term demand expectations for DRAM and NAND bit growth drop from 20% and 40%, respectively, to something meaningfully less than that. Ironically, Micron’s valuation (multiple of earnings) is likely to go up in this scenario, but not enough to compensate for how much earnings are likely to decline.
I think the best case scenario is dramatically more likely than the worst case, but there are many potential scenarios in the middle that could come to pass.
One last comment in this area. The way China is behaving here has got to be a serious warning to all other tech companies working in and around China and even Taiwan. While the Korean semi companies may benefit in the short-term, longer term they and every other semi company should be thinking very carefully about how China’s behavior could be easily turned on them. This should work to the detriment, over the long-term, to how foreign companies cooperate with/invest in China.
Schiefelbein: Will Moore’s Law continue to hold true over the medium term future?
Gregg: Moore’s law is already breaking down in a big way in DRAM. The technological complexity is such that the major DRAM manufacturers are spending a lot more to achieve less gain in terms of shrinks. This is one of the primary reasons why DRAM is becoming such an attractive product. Demand growth has stayed robust (driven largely by cloud/big data and more DRAM content in mobile devices), but obtaining bit growth through shrinks is not even close to keeping up. In NAND, currently bit growth is staying very much on track with Moore’s Law, but not because of shrinking cell sizes. Much higher bits per chip are being achieved by vertical scaling and splitting memory cells into fourths from thirds which is materially increasing memory capacity per cell. In logic, major semis are having difficulties (Intel most notably at the 10nm node) ramping at smaller and smaller transistor sizes. Process complexity is increasing and some of the newest tools needed to manufacture at smaller transistors sizes (5nm and 7nm) can’t process nearly as many wafers per hours as traditional deposition and etch tools can at 10nm and above.
Schiefelbein: What is your take on the recent New York Times article exposing the “heist” of IP from Micron’s Taiwan office? What impact could an insurgent state-subsidized Chinese memory industry have on Micron?
Gregg: The NYTimes article was good, but the WSJ article added even more to the story so should be read as well. And most important is to read the press release issued by Micron recently. The question is even with what China has mis-appopriated and with all the talent it has hired away at 3-5x prior salaries at competitors (like Micron), how quickly will China become a real competitor in this market. There is also the question of what remedies will foreign companies and foreign states (like the US) enact on China for its illicit activities. Big picture, most industry observers believe that China will not have a major impact on these markets for a number of years. The WFE companies do expect demand for their tools to ramp, but they have been very restrained in their expectations for how quickly domestic Chinese companies will become major competitors at leading edge nodes in memory and logic semis. As long as secular growth trends continue for DRAM and NAND, there should be plenty of demand to go around.
Schiefelbein: How does your background as the CEO of OTO Technologies play into your investment philosophy?
Gregg: Interesting question. My time leading Oto helped me develop a broad understanding of how IP is developed, acquired and monetized in the market place. It was a great complement to the work I did as an investment banker at banks like Goldman Sachs, Deutsche Morgan Grenfell, DLJ and Credit Suisse. That work ran the gamut from credit analysis, to capital markets, to advisory to C-suite executives on how to best drive shareholder value. My time leading Oto helped cement my broader view about looking for “fat pitches” and investments with the following characteristics:
- High ROIC, ROE, RONA business
- Strong current or emerging cash flow or profitability underlying the business
- Expectations priced into an investment that are moderate or preferably lackluster that provide significant upside to the extent that the market is likely to come around to revisiting those expectations higher
- Attractive valuation ideally vs. both peers and the broader market
- Management evidencing a smart history of capital allocation
- Attractive growth characteristics
- Companies that have the potential to significantly re-rate that are going through either strong secular change or are still early in a notably improving cycle
Schiefelbein: What is the biggest risk to your thesis? Why?
Gregg: Similarly to how all people die from the same cause (lack of oxygen to the brain), the biggest risk for memory firms time and again is when the supply/demand relationship develops in an unforeseen manner. If demand drops precipitously for both DRAM and NAND just when significant new capacity needs to be digested by the market, memory prices drop like a stone as do stock prices. There are numerous secular trends that are stoking robust demand and a few major forces (significantly increasing capital intensity, technological complexity and more disciplined industry structure/behavior) which are providing more disciplined supply growth. Based on Micron’s rock bottom P/E and EV/EBITDA, the market is clearly pricing in that the best has already transpired and that some, seriously negative (for the memory semi manufacturers) supply/demand imbalance is likely to come in the near term. I submit that the market is wrong. And even some notable recent bears like Tim Arcuri at UBS have started changing their tune and become more positive on Micron.
Schiefelbein: Where else do you see value in the market today? Where else does your fund focus?
Gregg: Our core strategy (“The Capital Appreciation Strategy”) is heavily weighted in tech and financials at the moment. Tech is split between some blue chip GARP investments (like Facebook, Google and Apple) and then a number of WFE, semi and capital equipment stocks that are all suffering at the moment from these trade fears but that I believe will be a home run over the next 12-24 months as some of these clouds are likely to dissipate. What I like about large/mid-cap tech we’re invested in right now is you have secular growth drivers, typically significant excess cash balances (that these companies are just now, due to tax reform and better ability to repatriate cash, looking to use proactively to create major shareholder value), enormous profitability and undemanding valuations relative to the broader market.
On the financials side, I like insurance (AIG - a great turnaround story with new, competent management in the saddle and Voya – now a clean story with the best forward EPS growth, highest ROE in its peer group and smart acquisition candidate) given where we are in this economic and rate cycle. I also like AerCap, which I submit is a massively undervalued commercial aircraft leasing company (that also would be a smart takeout candidate (maybe even an LBO candidate)). AerCap has been effectively doing a “rolling lbo” for more than two years now typically buying back 3-4%+ of its stock every quarter. They are in an awesome sector with high demand for their aircraft, much higher margins than typical lenders (and their “loans” are effectively secured) and with the strong, secular growth in passenger miles flown each year. AerCap trades for less than 8x anticipated earnings and at a significant discount to the market value of its fleet of aircraft.
Overall, it’s been a volatile year-to-date for the Capital Appreciation Strategy (unlike 2017 when our strategy was up ~32% after fees), but I believe we’re well positioned for the latter half of 2018 and over the next 12-24 months, absent a recession hitting (which I don’t anticipate over that period).
Disclaimers: Past performance is no guarantee of future results. All information is gathered from sources believed to be accurate, but the accuracy of the information cannot be guaranteed. A number of points made are matter of opinion and opinions are subject to change without notice.