Why HBM is the Hottest Thing in Memory

How a new memory technology is transforming the industry. Plus further ideas behind the paywall.

Hello everyone. There’s been quite an inflow of subscribers and new followers, so I thought I would refresh some ideas for everyone. At a high level, I want to talk about my philosophy about investing in semiconductors and then talk about ideas.

Disclaimer: Please note that this newsletter is not investment advice. The information provided is for informational purposes only and should not be considered investment advice. You should consult with a financial advisor or other professional to determine your investment goals and risk tolerance. Any investment decisions you make are solely your responsibility.

---

Investing in semiconductors can be pretty simple if you let it be. At a high level, I believe you want to invest in the secular at a decent price or invest in places where there are unwarranted dislocations. Sometimes the entire ecosystem says one thing and the stocks say another. Usually, the ecosystem is right.

In the past, I’ve probably done my best in the latter for my “picks,” and I usually continue to write positively about the former. It’s hard to choose when it's an excellent time to buy a secular stock, but it seems impossible for companies like BESI or NVDA to time the buys, rather than hold these high quality and secular stocks.

Where I think I excel at semiconductors is the second category of pointing out companies that don’t make sense given the entire industry context. To a certain extent, a bit of a long-short mindset is always applied to these themes. For example, my bearishness on Microchip is relative, not absolute. My best hits are often relatively fantastic performers, like Rambus, one of the few positive stocks for 2022 in semiconductors.

Anyway, I think that, for the most part, I have a smattering of relative ideas in today’s post. And the final set of ideas is a secular play, albeit it’s risen quite a bit. I will talk about HBM in front of the paywall and then the rest of the ideas behind a paywall. HBM’s dynamic in the market must be discussed, so I thought everyone should read it.

HBM and Memory

I’ve been optimistic about HBM and SK Hynix and wrote about that in my 2024 outlook.

The clear winner here is SK Hynix. So, while I don’t prefer memory stocks, if you had to own one, own the company that will likely continue to have the largest share of HBM. Especially because the other players have minimal share (Micron) or are a conglomerate (Samsung). Sk Hynix is the best pure-play.

But as time has passed, I’ve decided it’s worth dumpster diving in Memory companies. I think Micron is a bit stupid right now, but I particularly like the leader in HBM, SK Hynix. Hell, even Micron will likely benefit from this cycle from HBM.

HBM is the thing that is different this time. Both the demand and supply aspects of HBM is making me bullish the entire ecosystem, and I think there is irony that after the deepest and worst memory cycle in recent years, HBM will lead to the strongest cycle yet. But to really drill this analogy home, I’m going to try to use Oil and Gas producers as an analogy. Right now, it’s all about supply and demand, and that’s the best industry for this analysis.

Supply Growth and HBM

Memory is closest to a pure commodity of all the markets in semiconductors. Despite further consolidation, memory still goes through boom bust cycles given that it’s a finished product that requires a lot of capex, and the incumbents can always make more if they spend more. Memory is much like energy because you can spend more capex, and then increase supply, and while there is secular demand for bits, if you increase supply to quickly, you lead to oversupply. And even like energy, there are separate types of commodities.

I will call NAND natural gas and DRAM crude oil to set up this analogy. HBM in this case is the brand new SuperOil+, and let’s consider it a superset of oil, and is drastic demand right now.

Let’s pretend that oil and gas producers magically found a new type of energy that we mentioned called SuperOil+. They can repurpose and buy equipment they are already familiar with to drill this SuperOil+, but it will take 2x as much equipment to yield the same volume of SuperOil+. However, SuperOil+ costs 5x as much as oil. Additionally, SuperOil+ is in insatiable demand, and as much SuperOil+ as you can drill will be bought at least for the next few years.

So what will you do? If you’re an energy producer, you will likely put every dollar of capex towards SuperOil+ because the returns are better and pull back capex from Oil and Gas spending, which ironically lowers supply growth. Now the supply and demand curves of oil and gas are not impacted by SuperOil+, so supply growth going down will help price stabilize even faster. This means that SuperOil+ is a windfall as a new market, and a windfall for the traditional oil and gas market. The entire ecosystem will be more profitable because supply growth in the traditional oil and gas ecosystem is going down. That’s exactly what’s happening today, except replace SuperOil+ with HBM.

HBM has the unique dynamic of costing a lot more than traditional DRAM, and the dynamic 2x higher capex is driven by the twice as large die size for HBM. So in order to increase volume in absolute terms, it will take a lot of capex.

Now there’s some context. Memory companies are coming out of a historic downturn and have only just recently stopped burning cash, so they are constrainted in their capex budgets. They will shift what cash they generate to HBM capacity, which will keep supply growth in DRAM and NAND low. This is the best possible case for memory companies.

HBM will kick off quite the boom for memory companies; I think it’s just beginning. We will see new cycle highs in a few years, and I feel it makes sense to bet on the leader, in this case, SK Hynix. While it’s domiciled in South Korea, it’s trading at a discount to Micron, which has started to rally on optimism for HBM3E. Please keep it simple and bet on #1, especially if it trades at a discount to Micron. Samsung is not the purest play given its large logic and smartphone businesses.

Until we run out of demand, it’s time to drill baby drill for the memory companies. That will be a period of at least a year or two and is a truly historic moment for the memory companies. It’s going to be an interesting cycle, to say the least.

This free post is dedicated to RaisingTheBAR47 on twitter, who taught me all companies can be descibed as oil companies.

For further ideas, please subscribe, as the rest of this post is paywalled. One subset of ideas is where the ecosystem and stocks are disagreeing, the other is chasing AI to it’s logical conclusion.