Semiconductor Outlook 2023: Green Shoots
It's the long Winter. Will Spring ever come?
2022 was historic. There’s no other way to describe it. The huge 2020 semiconductor shortage turned into a glut, a tale as old as time. Last year I wrote about the cyclical becoming secular, and that was about as wrong as you could be. But besides the cycle, the other big story this year was China. But let’s review what happened this year so far.
Early in the year, the story was still about how the cycle was ripping higher, and TSMC was consistently crushing estimates and ordering more equipment. WFE looked like it was marching well on its way to $100 billion, but then the cycle started to show its full fury. In Q4 2021, there was insufficient inventory, so overordering and supply chain issues inflated the channel inventory for almost all consumer goods. Given how hot the economy was running, inflation was running rampant.
The world finally reopened for good (ex-China, more on that later), and the reopening shifted spending from goods to services, and thus the inventory ordered for the Covid splurge unwound. It wasn’t just semiconductors that had too much inventory. Shoes, furniture, to everything that benefited from the work-from-home boom had too much stuff. And that wasn’t the entire story.
In this environment, inflation was spiraling higher, and Jerome Powell decisively started raising rates to stop inflation from spiraling into stagflation. This steep rate hike sent global shockwaves, as dollar inflows due to higher rates created a strong dollar and crushed exports from the US.
I like the Fed Funds rate chart because 2022 essentially is a straight line upwards. Thanks,! (You should follow Joey, he’s a sharp guy!)
Elsewhere Energy crushed Europe and further fueled inflation in non-core goods higher. That part of the story was already happening, but the Ukraine war exacerbated a supply-driven shock to the price of oil. All of these factors would soften the economy, especially the forward outlook for companies. By March, it was clear that China’s market, in particular, looked oversupplied and that the coming crypto proof of stake would start impacting Nvidia. But all of this was amid TSMC beating monthly revenue each month and accelerating.
By the first earnings season, I was beginning to talk about the macro much more. And by the time it was clear the cycle was beginning, I started to look back into history in the 1980s and 1990s for more insight. It seemed like we were in a classic inventory cycle as we got to June. In September, Rich Templeton started to talk about this cycle akin to the 1995 cycle, one of the hardest in recent times for memory players.
In October, the hammer dropped on export controls, and the China-US trade war got hot. This was not good for semiconductors, as China is ~30% of global demand. Oh, and the rolling lockdowns impacted Chinese demand all year. So now it’s the end of the year, and SOXX is down 35%.
Stepping back, it looks like semiconductor sales increased ~4% this year over 2021, and WFE increased by high single digits over 2021. The marked deceleration means that semiconductor sales have been meaningfully down in the second half of this year, and the stocks reflect it. Semiconductors were one of the worst performers until October and have meaningfully rallied since then. They are slightly ahead of the software and communications indices but lag technology broadly.
It’s been a rough year, and semiconductors underperformed broader indices. It’s been a historical cycle, yet numbers are just now starting to go negative. It’s looking rough, but I have some belief that the future is going to be alright.
The biggest thing that is getting me excited is that it looks like China is doing a swift about-face in its covid policy. The near-term numbers will be horrific with the mass wave of Covid in China, but a quick reopening could improve an economy running with slack. This is good for semiconductor demand, especially in the most bombed-out section of Chinese handsets.
Additionally, I think the inventory cycle may end sometime in the 1st half of 2023. Refer to this piece where I talk about the inventory cycle historically. I am increasingly confident in this, especially if China’s economy improves.
So between an inventory bottom, a hopeful end to rate hikes, and decelerating inflation, I think there’s a lot to be excited about in 2023. I have no idea what the shape of recovery looks like, but I can see the green shoots from here.
So now I want to tell this year in a different format, numbers. Let’s talk about single stock performance and valuations. As well as take a walk through each end market and talk about my outlook.
The Year in Numbers
Here are some of the companies I track. Stock performance wasn’t great.
I want to take a second and highlight RMBS, one of the few positive stocks year to date and one of the companies I pounded the table on hardest. I am proud of that one, and between the podcast and the post, its relative outperformance in a tough year is amazing. It sure helped me!
But things were not all sanguine. Many companies I wrote positively about or even negatively about (KLIC, I am looking at you) did poorly. But I want to point out a very simple observation, many of this year’s greatest losers were last year’s greatest winners. I bring this up because I noticed this last year (surprise!), and I think that bombed-out semiconductor companies with decent business are a great pond to fish in. Below is a chart of 2022 returns vs. 2021 returns.
Most companies are still up over two years, but a few are now down on the 2-year return mark. Here are the 2021 and 2022 returns stacked.
Now let’s compare valuations as well - starting and ending. Almost every company’s multiple compressed this year - but much less than the actual stocks fell. EPS estimates increased for a few companies, namely ASML, TSM, ADI, and TXN. Semicap earnings fell less than expected (ex-LRCX), and INTC and NVDA estimates plunged.
Taking a more granular look, we can look at the EV/EBITDA multiple of the universe right now. I think the valuations, for the most part, make sense and are a bit less ebullient than in the past. The upper end of valuations (MPWR, ASML, NVDA, LSCC, AMBA, WOLF) still feels hefty, but this valuation sheet feels much less insane than a year ago.
Companies with low multiples, poor 2022 returns, and strong businesses will likely be the interesting stocks of tomorrow. Now let’s walk through each specific end market and see how the companies did.