I like to remind people that “when it comes to finance, most things are pretty obvious; it’s the timing that’s hard…”
Through nearly three decades of investing, this concept has constantly proven itself out. Sometimes I really do get it wrong, but usually, I’m just early. As a short, this can be incredibly painful; though as a long, it’s mostly just frustrating to lock up capital that’s going nowhere. As you can guess, this is the reason I rarely short. That doesn’t stop me from calling balls and strikes. I’ve done this in many prior situations (here and here). Sure, I was often early, but the fever dream eventually broke, and the securities plummeted back to earth. I bring this all up as I’m stupefied by the mania in semiconductors.
To frame this situation, it’s important to remember that semis are incredibly cyclical with multi-year boom and bust cycles. They’re unusually capital intensive as fabs cost billions to assemble, but unlike many other capital-intensive industries, they also face technological obsolescence. They’re the ultimate sh*tco sector. I literally cannot think of a worse sector to invest in—though I freely admit that every so often, they launch themselves on a super-cycle.
We saw one such super-cycle peak, just as I was coming of age as an investor in 2000. The road down was brutally painful for investors—supply continued ramping as fabs came online; despite orders vaporizing as the tech bubble ended and companies that triple-ordered went bankrupt. At the nadir, which lasted more than a decade, many of the current high-flyers resorted to toxic financings to simply keep the lights on. We’re now into the second such mania of my career.
I get it, AI is hot and you don’t have AI without semis. The buyers seem to be more financially stable than the dot com and fiber players, and the profits are real. Anyone who nailed this on the long side deserves respect. I missed it. However, I also know enough to know that semis are still highly cyclical, and after the peak, there’s likely to be an incredibly painful down-cycle. As a result, everyone invested in semis should be asking themselves a few hard questions;
-How high will peak margins get??
-How long will the cycle last until new supply gluts pricing??
-How much will my favorite semi earn before the glut comes??
-How does that level of accumulated after-tax earnings from now until the glut takes pricing negative, compare to the current Enterprise Value??
Everything else is irrelevant. The fabs will mostly be write-offs (for much of my career, these things traded at large discounts to stated PP&E). The R&D is a recurring expense. The equity dilution is very real. So, what’s left over is simply the after-tax cash flow from today until the day that the glut happens, because if I know anything about highly cyclical sh*tco industries, it’s that once the glut comes, these things will bleed capital for many years into the future. Do these semis trade at a premium or discount to that value?? I don’t have the answer myself, as noted, I avoid technology. However, if you aren’t constantly asking yourself this question, you’re not investing—you’re playing momentum.
But Kuppy, where does the supply come from if everyone is a disciplined operator??
Have you heard of China??
As my good friend Louis-Vincent Gave always reminds me, “when China enters a room, profits walk out.” I can think of dozens of such situations. Everything from steel to aluminum, from chemicals to solar panels, and shipbuilding to electric vehicles. China has historically followed a recurring pattern—they show up, they use cheap funding and they glut the industry. Then once everyone’s margins (including their own) are deeply negative, they ramp capacity to many times global demand, before they break international players, yet their utilization levels continue to drop as they keep building even more capacity. It really is the runaway assembly line. I’ve seen this over and over, in dozens of industries. Over time, I’ve learned to get out of the way when China is coming for a sector, because they simply will not stop.
Interestingly, China has repeatedly told the world that they plan to follow this playbook in semis, yet the world seems oblivious to the danger. Some seem to doubt that China has the technological knowhow (they said the same about EVs, but have you driven the newest BYD??). Some say that the West will never transfer the technology—have you seen China respect patents?? Some just think that semis are somehow special. I can assure you they’re not any more special than solar panels were. The Chinese are coming, and the glut will be legendary. In my opinion, no matter how many datacenters get built, they will NEVER be able to absorb the coming supply.
Then again, a lot of profits will get banked before China pushes profits in semis negative, but that time is coming. Maybe it’s 1 year away and maybe it’s 5, but it’s coming. If you haven’t thought deep and hard about when and how much you’ll earn before then, this should serve as your wakeup call. Once China starts to ramp, they won’t stop.
Remember, when it comes to finance, most things are pretty obvious; it’s the timing that’s hard…
The New Levels…