Blessed with poor memory
The best way to a happy life is to be forgiving to other fellow souls, the second best choice is to have a poor memory. There are a few exceptions, the exemplar one in the memory space is a company in Idaho, called Micron technologies.
We can all passionately debate that teaching sand to do math was generally a bad idea and humanity was better off without this invention. But here we are in the 21st century and silicon can perform 3 key functions for us: it can process (think CPUs, GPUs, TPUs), store (think SDD, HDD, blobs) and something in between is a working slate that we call memory. (Heard of RAM?)
Manufacturers of xPUs (Intel, AMD, NVDIA, et. al.), storage (Seagate, Western Digital, etc.) have all had their glory days and some of these shot to 10x in short time periods. The hope is to find something that shows similar success with odds stacked up in favor and Micron is right up there in the probability distribution.
We will look only at opportunities with a decade horizon as we don’t believe in betting on horse races. We will only consider things that have a very high chance of at least quadrupling in that timeframe. That would give us returns of about ~14% return pre-tax, ~10% post-tax, excluding dividends, beating most options when evaluated consistently against a decade long time window (excluding digital snake oil ledgers that may or may-not be immutable).
Why memory? Two decades back this was the stupidest idea. There were about 48 players trying to build memory. Building the next chip required hundreds of millions in R&D investments as everyone was trying to squeeze in more transistors (building blocks that convert sand into a mini-mathematicians) in a silicon wafer and the price was decided by your dumbest competitor trying to clear of inventory of their low performing chips. But every modern day device needs memory.
Why now? George Moore, the founder of Intel, said that the number of transistors that could fit in the same sized silicon wafer would double every couple or so years. This was true until most recently, we are now hitting the limits of physics where we have transistors that went from something the size of a light bulb to present day where they are as thick as the hair of DNA and will soon approach the size of a few atoms. Each iteration, pushed about half the competition out. 48 players became 24; 24 dwindled to be 12; 12 waned to 6; 6 halved to 3. We are now left with Samsung, SK Hynix and Micron. (It is important to note that Intel, AMD, NVDIA and others DO NOT make memory.)
What formerly was a cyclical commodity business with huge waves of pricing driven by demand and supply is now transitioning into a space that will see rational pricing driven by a mature oligopoly(, a condition where few dominant players rule the market).
Taking a short and rare pivot to geopolitics, while it is sad to see what is happening in Ukraine 🇺🇦 or looking back at the pandemic, USA 🇺🇸 has woken up to the fact that they cannot heavily rely exclusively on a robust global supply chain or strong trade ties to advance their economic interests. Additionally, the heavy tech investments by other countries is challenging America’s tech supremacy. Which made the Congress bring all major semiconductor companies including Micron and present their case that led to the passing of a semiconductor bill which will make $39 billion dollars available to semiconductor manufacturers in ‘Murica!
This will grow together with other industries including EVs, data centers (cloud), 5G, IoT, gaming and even coin mining if that fad ever picks up again.
Why Micron? These people make the best in class memory, that is protected by over 50k patents, manufcatured right here, in the land of the brave, home of the free. When any bigh tech company sets up a datacenter (a piece of cloud on earth), 30% of the cost goes to the memory guys. Think of it as a tax on all your shopping, all your steaming, or every-time you touch any piece of tech.
Micron is cheap based on all your traditional (single-digit) fundamentals multiples relative to most US stocks and stocks in the semiconductor industry, has high cash reserves, high consistently growing revenues, has a wide moat (or high barriers to entry protected by a ), top quality management (CEO was former co-founder of SanDisk), has committed to 150+B in R&D, pays dividends and buys backs shares. All these stars rarely line up.
Why should we believe you? You shouldn’t actually. Despite the fact a random person on the internet who claims to own one part million of this less known company with his sizable networth deployed, statistically you were better off indexing in a low cost market fund. It has worked in the past and with a lot of people parking money there twice a month via retirement savings, may give modest returns. A word to the wise though, much like fashion, investing loses it’s charm when everyone does the same things.
Now what? Enjoy your Sunday or whatever day you stumble upon here. A decade back I was just transitioning from being a mushroom farmer to a physics teacher; A decade hence I hope to offer capital allocation services. It would be great to start a 0 fees-performance based fund then. I will provide you with:
a) An auditable track record of my ideas and performance with the timeline and % allocations (1/6th my current networth is deployed in Micron)
b) Simple easy to consume analysis of the business thesis that looks at the story as most analysis are inundated with numbers.
c) An option that is compelling measured against most choices spanning over a span of a decade.
This writeup with either serve as a good marketing content for a brochure or remind me of some of my blind spots. Either case, it helps to put these down if you are blessed with a poor memory. 😁
What is your best idea? LMK!
More from bubbykin here. You may also like some of these popular posts:
Investing in high growth economies
