No. 32: Christmas Inflation and Investing Thoughts, Micron Earnings Highlights, Semiconductors and Software Development
This is my last post of 2021. Thank you to all the readers, supporters, critics and sparring partners through the year. I hope at least some of the articles have led to insights and better decisions (or bad ones avoided).
I leave this year with a couple of thoughts before going through the customary investing/business snippets.
Luck and Skill
It’s hard to determine whether someone is skilful or lucky. In addition, what has been interesting for me (based on the massive crypto and FAANG stock run) is that early investors in things with massive potential scale can succeed without much intelligence. One can simply buy a basket of many potentially massive ideas, and the magnitude of any success will outweigh the 90% of ideas that might fail. For instance, putting 3% of your net worth in Google and another 3% in Bitcoin, with the rest in Berkshire Hathaway wouldn’t have yielded bad results even if the speculative part went to zero. But if any one of those ideas succeeded at scale, it could have been massive. Of course, hindsight is 20/20 – but you get the point.
Therefore, the debate is not just between luck and skill anymore. It’s also about how early one can join the party – even with a small % of net worth – David Gardner style. And as we’ve seen, the last decade has been kind to early investors of nascent tech businesses.
Assuming that this is true for the next 20 years, this leads to a kind of game within a game. A bit like venture capital. How can you get early ideas with scale passing by you? Which contacts do you need to cultivate? Can you do simple things like rearrange your Twitter feed? Investing behind idea flow can yield massive results – even before you make a single stock pick.
That’s not to say skill doesn’t play a part. I notice that I’m increasingly uncomfortable with being a jack-of-all-trades, flitting from banks one day to oil stocks another day. So, for my larger positions I’ll hopefully stick to what I know.
Inflation
Many consumer businesses I follow have experienced margin compression (excluding mix changes). Some compression results in ad budget reductions unless CPG companies can put through price increases with retailers. This means ad supported revenue companies may take a hit from CPG customers in the short term. On the other hand – advertising mix will change to advertisers who are gushing with cash. For example, I’ve noticed a tech company ad blitz everywhere I look (even outdoor).
Semiconductor inflation and shortages have been rampant. As a result, chip customers like Apple and NVIDIA now make upfront payments to TSMC to secure supply. We can extrapolate that this means additional capital requirements to stay in the industry. Weaker players, or players whose customers don’t front up with cash might get knocked out if they can’t produce the upfront payments (as I also wrote about in an auto industry article). In fact, this goes for any other industry experiencing shortages. Supply goes to those who can pay up and guarantee massive scale.
Of course, businesses won’t sit still. They will keep finding ways to offset margin compression. From recent conversations, and also from the recent AWS:reInvent conference, it looks like industrial companies are seeing massive ROI from cloud spend.
Therefore, the industrial cloud ecosystem will probably experience a long revenue tailwind. If industry gets increasingly automated and data gets easier to collect and analyse, utilization and efficiency of manufacturing might increase, driving margins back to where they were pre-COVID.
Supply chain vulnerabilities have become obvious through COVID (and Brexit in the EU). Wage inflation has become rampant. Therefore, the business case for warehouse and handling automation has never been better. The secular truck driver shortage has temporarily been solved with higher wages and signing bonuses. But over time, the industry will need to address this with several technologies, network designs, and innovations.
Let’s see how these issues unfold in 2022. There are so many moving parts that making a prediction and believing in it is probably not wise. But imagining potential scenarios is nevertheless necessary to form investment judgements.
I wish you all a Merry Christmas and Happy New Year! Subscribe, share and keep reading in 2022.
Micron Earnings
Micron absolutely crushed it – according to the market response sending shares up 10% straight after the earnings release. There were a few things that I highlighted from the earnings call which I find of interest.
Firstly, the fact that memory and storage spend has outpaced the rest of the semiconductor industry. I have spoken about this trend before here, and it looks set to continue given that most computing needs of the future are very memory-hungry:
Memory and storage revenue has outpaced the rest of the semiconductor industry over the last 2 decades, and we expect this trend to continue for years to come, thanks to AI, 5G and EV adoption. In addition, the build-out of immersive virtual worlds often referred to as the Metaverse will offer even more opportunity due to the intensive use of significant memory and storage in these applications.
The mobile upgrade cycle continues to provide the memory industry with a growing revenue stream. And, given that each new model contains more DRAM and NAND than the previous model, this secular upgrade cycle also looks set to continue for the foreseeable future:
Recent 5G phones feature more than 50% higher DRAM and double the NAND content versus 4Gphones. 5G smartphone sales are forecast to exceed 500 million units in calendar year ‘21, with 700 million units forecast for calendar year ‘22. We expect mobile content to continue increasing as 5G phones benefit from further innovation in 5G-enabled applications.
If Micron sells 40% more units in 2022 than in 2021, combined with 50% DRAM content growth per 5G handset, this equates to 20% overall 5G bit shipment growth despite overall mobile penetration staying flat across developed markets. The mobile upgrade cycle is very powerful for Micron despite being a “mature” market.
There’s some consensus in the semiconductor industry that automobiles are the next secular driver of overall semiconductor content – as I wrote about in this post. However, Micron’s earnings transcript talks about the opportunity for DRAM and NAND based on the latest high-end vehicles:
New EVs are becoming like a data center on wheels, and we are already seeing examples of 2022 model year EVs supporting Level 3 autonomous capability with over 140 gigabytes of DRAM and also examples with over 1 terabyte of NAND. In addition to continued content growth, we expect calendar year '22 auto unit production to increase as non-memory component shortages ease. We entered into a new supply agreement with UMC to improve our ability to support our automotive customers with NAND solutions as market demand strengthens in calendar year '22.
In comparison, my 2013 Toyota hybrid probably has almost zero DRAM and very little NAND. The most complex chip in my car is probably in the little touch screen which is used to turn on the radio. Contrast that to an L2 self-driving EV which requires all kinds of systems – and are increasingly becoming a larger portion of new vehicle sales.
The total amount of cars is also interesting to ponder. Here, management of Micron compares the total number of cars to the number of data centers to illustrate how powerful and secular this growth engine could be:
…And actually, in automotive, if you think about it, we have 90 million automobiles versus servers at about 15million. So 6x of automobiles with increasing content. So of course, this will be a strong demand driver, not just for 2022. I mean, beyond that as well. So the demand trends are secular here in nature.
I’m very excited about the opportunity for DRAM and NAND growth in years to come. Of course, technology changes fast, so Micron can’t really rest on its laurels. However, if the customer ROI of DRAM and NAND usage stays high, there will be little incentive for the industry to adopt alternative competing technologies at a rapid clip (if they can first find them, that is).
Semiconductors and Software Development
Everyone knows there are huge secular tailwinds to the semiconductor industry. However, at some point, if tech scaling challenges and shortages persist, ROI of software efficiency becomes attractive.
Software optimization salaries have doubled every year over the past few years. If efficient software means you can avoid writing billion-dollar cheques to the semi oligopoly, then it pays to hire great engineers who can do just that and pay them ridiculous amounts.
Loved the first part about luck and skill!