Unit Cost: $42
MU
Pabrai $38-$39
“It was our kind of business. It’s one you can understand, I may not understand all of the products that they sell and I may not understand what the customers who buy the products do with them, but I do understand the middleman’s role.” -Warren Buffett on Tech Data
(time of write up was @ $33)
(2019) My MU investment is a pure clone play. David Tepper(27%), Mohnish Pabrai(10%). It is out of my circle of competence. There are many aspects of the industry and company that I don’t feel confident in. I couldn’t go into detail about the technology of 1Z, NVMe, or eMMC and I am violating the cardinal rule of staying within my circle of competence. However, the technology hasn’t been replaceable these past few decades and currently no signs of changes in the near future. The economics seem straightforward for this cyclical company. It is an oligopoly and it’s past cyclical nature is behind it with the limitation of material physics, it has a net positive cash balance, has succulent margins, with plans to return cash(2019) to investors via buybacks(20%) and dividends. It has multiple tailwinds as we continue to embed technology further into our lives. In addition it is not able to be profitable during a down cycle and has the ability to blunt costs and preserve its margins during the down cycle by decreasing cost per bits, opex and capex as needed. As long as these metrics are preserved, I believe the company should survive the current down cycle and be a fine investment due to its low valuation.
Summary
With its low valuation, Micron is a way to capitalize on the technology revolution without having to decide who the winner will be. Being long Micron is being long technological advancement. Micron is part of an oligopoly that supplies the new fast growing tech companies with memory chips to the emerging need for servers, cloud, smart cars, gaming consoles, advent of 5g, laptops, pc’s etc. The business will thrive on the basis that it will be a tax on the technology industry as a whole as we enter this new age of tech. It is not hard to imagine that the current high growth rate of technology will continue for decades.
Business
“Micron makes memory; the chips that allow a computing system to store information (either temporarily or on a longer-term basis). Memory chips are quite different from the most commonly thought of semiconductors: logic chips, which have the job of processing data. Memory chips don’t process data, instead, their job is to feed data into processing chips, similar to a hopper in a factory, or a storage unit. Without memory, computers can’t function, and human productivity stalls. For logic chips, most companies create their own designs but outsource their production to Taiwan Semiconductor. In memory chips, production is vertically integrated; the same companies that design the chips also make the chips.” [1]
Micron makes memory chips in the form of DRAM and NAND. Revenue contributions are approximately 65/35% give or take +-5% respectively. DRAM is an attractive business, NAND is not. They are the 3rd largest player in DRAM, DRAM chips act as the main memory in PCs, smartphones and other devices, working closely with central processing units. NAND flash provides longer-term data storage. Currently the oligopoly consists of three DRAM players(Samsung, SK Hynix, Micron) and five NAND players(Samsung, SK Hynix, Micron, Western Digital-Toshiba, Intel). Samsung has the largest market share but also consumes its own product.The industry is currently in a downturn due to supply outstripping demand and China trade issues.(2019) They are currently improving their already superior cost competitiveness which will help to navigate through the downturn. Micron will continue to be cyclical, but not how it was in the past. Each peak and trough should continue to be higher than the next showing lumpy but continued growth. Also, they should be profitable during downturns which is different from how they have performed in the past.
Revenue is segmented into four business units, revenue contribution always subject to change
- Cloud and network business unit(CNBU) 51%
- Mobile(MBU) 23%
- Embedded(EBU) auto(full self driving(FSD) and others 12%. storage Business Unit(SBU) – solid state drives(laptops etc) and others 15%
Micron is a capital intensive business, but despite this, it still has high owner earnings. It is currently in a net cash position with plans to return $10B(20% of market cap) in the form of buybacks. (2018)
MU: Profit Margin: 37%
Operating Margin: 41%
Return on assets: 16%
Return on equity: 31%
SK Hynix: Profit Margin: 28%
Operating Margin: 37%
Return on assets: 14%
Return on equity: 22%
Change in Landscape and what it means
There used to be 30-40 players in the game, but it has been reduced to three(DRAM) and five(NAND) through extinction, competition and acquisitions. In the past, the industry was nascent and rapid improvement occurred at the clip of 40% a year in terms of the density of information that you could fit on a sq nanometer of wafer and cost of improvement where you would get more yield from a meter of silicon. Because the technology changed so fast, it meant that if you didn’t keep up in a matter of quarters that the competition would create a better and cheaper product rendering the inventory obsolete. You would have to get rid of it quickly in a matter of weeks / months at fire sale prices as dramatic as $3 to $0.50cents. This would lead to losses during troughs. Today, the pace of improvements have dramatically slowed with the current pace of innovation slowing to about 4-5%/year. We have hit the limits of what can be achieved in material physics. We already have billions of transistors on a nanometer of wafer, we physically can’t fit that much more on at a high pace, material physics prevents this from growing quickly. The cost to get that slight improvement is so much higher, so new products cost so much more but aren’t that much better so clients are inclined to stay with the previous model. This allows Micron to increase its shelf rather than to have to quickly get rid of it at fire sale prices because their product was obsolete. The preservation of prices allows it to prevent it from money losing troughs. 100 day of shelf life used to be the leading indicator for downturns and the collapse of profit. The company would lose money because it meant they had to unload their inventory at a loss. In 2019, they achieved the unprecedented and made money during the downturn because they were able to continue to preserve their prices and margins.This also results in better return on capital. A new DRAM fab cost up to $3-4B, return on capital is better than in the past bc it’s now an oligopoly and the 30-40 players aren’t eating your margins
Tailwinds
- Data Centers – GPU use in data centers. Newest nVidia chip uses 3x more DRAM than it’s previous model. AMD and nVidia both use DRAM. You don’t have to guess which one wins, you will ride the wave of both of their successes. (Servers are the most important vertical for DRAM – MU Chief Business Officer
- Mobile – 5g phones require 6x more DRAM than 4g phones.
- FSD – requires DRAM. A vehicle needs an estimated 300gb.We’re already at a shortage in the automotive industry needed to support simple things like heating the car, windshield wipers etc. There will be a seismic jump when FSD goes full steam. (fastest growing segment along with industrials
- Machine learning / Ai algorithms requires huge amounts of DRAM
Moat
1. Patents – 40K patents, years and years of protected innovation, and fabrications require an enormous amount of capital. DRAM is on it’s 5th iteration, you have to have the licenses / patents of the previous 4 iterations. The strength of the patents was tested in the court of law and Micron won the IP case. An engineer went to a Tawainese company w/ Micron’s secrets and Micron sued for infringement. They won and now the company has to pay licensing fees.
2. Capital & Technology – “Each of these memory makers has invested a tremendous amount of capital and expertise into their manufacturing footprint, and without many billions in investment and unencumbered intellectual property, it’s very difficult for a new competitor to emerge. Even though Micron is the third-largest player in DRAM, its capital expenditure budget for the upcoming year is expected to be $9 billion. Even if you have the money and the intellectual property to reach a cutting edge node, there is tremendous execution required to keep up with market leaders who have decades of experience and know-how. Without that ability, a new player will likely be left behind, chronically unprofitable, and with little to show for its massive investment. The advantages a Chinese company can bring to the memory business (low cost of labor and lower internal margin requirements) are not meaningful if the unit economics themselves aren’t competitive. In a scale chipmaking operation, the majority of the cost base isn’t due to labor but rather the remaining production costs in each chip (determined by manufacturing yield, technology process, and scale). There is a 20-30% cost difference per chip if you aren’t on-par with your competitors, and this is a 20-40% gross margin industry!”[1]
China
There are concerns of a high addition of competition from China. However, China is behind, full ramp up is not expected until 2020-2021. And when it happens, their inferior technology and manufacturing process will lead to volumes equal to 50% of the current manufacturers. There are also delays in manufacturing additions from Samsung and SK Hynix.(2018) They are currently not even close to full ramp.(2021)
China is importing $600B in US semiconductors a year and are highly incentivized to get out of it. However, the problem can’t be solved by just spending money. Enormous amounts of capital isn’t the only barrier to entry. The equipment required to make the chips are made by just 2-3 players in the world(machines such as etching and masking machines), ASML, Applied Materials and Zeiss, to name a few. There are US embargoes and sanctions preventing these companies from selling to Chinese companies. If they do, they will be prohibited from selling to US companies. 500 components are made by just 1-2 players. You can’t replicate the entire vertical chain and industry even if you’re China. For instance there’s a lens made only by Zeiss. In addition, they are 2-3 years behind on the technology and their quality sucks. At one point, China mandated the use of Chinese memory chips in their products. The companies couldn’t do it. The chips weren’t good enough. China’s leader in DRAM, ChangXin Memory Technology(CXMT) is shipping its first 19nm DRAM line, whereas Micron is working on 5nm: Yangtze Memory Technologies(YMTC) recently entered the 3D NAND market with 64 layer device, but has remained incapable of scaling production. Simply, China can’t buy the equipment, they don’t have the licenses, they’re behind on the technology and may never catch up and it is going to be expensive.
Valuation: The stock’s balance sheet, moat, and ability to increase its earnings are too great to be one of the cheapest stocks in the S&P 500. I have no target price, I just know that it won’t stay this cheap once the cycle revs back up and earnings increase. In 2018, analysts expected future earnings in 2020 to be $4.66(trough earnings-it ended up being $2.81 during covid) This would put it at PE 7 at trough earnings. Pabrai says he leaves the last 10% of his portfolio for a 4-5x in 5 years. He stated in his BC lecture that this could possibly be a compounder and will try to resist selling after a 4-5x. Micron can rerate cause it doesn’t lose money during troughs now and will have higher roc.
Waiting for $28(10% better than Pabrai, wow! this could also cost me huge opportunity, the #1 regret investors feel). Praying that this won’t come back to kick me in the ass. I just can’t help but remember Pabrai saying in an interview that 90% of the time, his investment goes down after he buys.
Pabrai on Micron
1. “I’m going to…really try to hold on, even when it gets to 4-5x the current value.”
2. “I do think, actually, Micron is a compounder, people don’t actually understand the business.”
3. Capex is not a deal killer….The memory players.. have high capex, but they also have high owner earnings”
4. “If Micron’s on your mind, then, at least for me, you have to wait some time.”
5. “ In fact, I’d be willing to wager that our largest holding, Micron Technology, will trounce a basket made up of the FAANGS and Tesla over the next decade.”2020
6. Made a bet on twitter in 2020 that the next 10 years, $MU will beat any stock of the twitter users choice.
7. “Every time a data center opens, 20-30% of the entire spending on the data center comes to the memory guys. It’s a very significant tax on data systems….it’s a high capex business and it goes through cyclicality and so on. I don’t think it’s a set and forget it kind of business. But it’s a nice business” 6/2021
8. 6/2021 Micron has more than doubled since we bought it three years ago. It is just getting warmed up
Here is a much more in depth write up, from someone who is more knowledgeable on the subject. https://www.valueinvestorsclub.com/idea/Micron_Technologies/3097958565#description
https://seekingalpha.com/article/4417750-micron-transformation-demands-valuation-rerating
https://lillianli.substack.com/p/china-semiconductors-and-the-push
The Rational Cloner 