Semiconductor Roundup - 1/28/2022
STM, KLAC, MKSI, Ampere Computing, Apple, Tesla, Silicon Motion, MACOM
This will be our second post of this type for the flurry of semiconductor earnings releases and calls that are happening now. In the first post, we covered Teradyne, Lam Research, Wolfspeed, Texas Instruments, UMC, Intel, Xilinx, Samsung, and MediaTek. This is required reading because the most interesting parts are what you can read through about other firms in the supply chain that sometimes are seemingly even unrelated. Some themes in the previous post will continue on, but others will not. In this post we will cover Tesla, Apple, KLAC, MKS Instruments, MACOM Technologies, Silicon Motion, and STMicroelectronics.
While Tesla isn’t purely a semiconductor company, they do have some really interesting semi ambitions as part of their larger overarching goals. We have talked about the semiconductor design ambitions in 3 separate articles last year, but I’m not here to talk about Tesla today. I just want to lighten the mood by quoting Elon from the recent earnings call.
Like that little chip that allows you to move your seat back and forth, that actually was a big problem. But a lot of these things are alleviating. I think there's some degree of the toilet paper problem as well where there was a toilet paper shortage during COVID. Obviously, it wasn't really certainly a tremendous enhanced need for ass wiping. It's just people panicked and got every paper product you could possibly wipe your ass with basically. And I wasn't sure, is this like a real thing or not? I actually took my kids to the H-E-B and Walmart in Texas to just confirm if it was real. Indeed, it was. And there's plenty of food and everything else but just nothing, no paper products. That didn't cause us [plenty].
An odd choice for people to panic about, yes, because these things are -- actually, if end of the world is coming, I think toilet paper is the least of your problems. So I think we saw just a lot of companies over-order chips and then buffer the chips. And so we should see, and we are seeing, alleviation in almost every area. But the output of the vehicle goes with the least lucky. What are the most problematic items in an entire car? There seems like at least 10,000 unique parts in the car, way more than that if you go further up the supply chain. And so which one is going to be the least lucky one this time? It's hard to say.
Okay you had your laugh? Now back to the serious business.
I don’t agree with Elon fully because inventories aren’t building that much anywhere across the auto ecosystem, there are some minor builds at component suppliers and auto OEMs, but nothing even close to the old levels of inventory let alone the mandates to have 1 year of inventories. The more interesting quotes relative to auto semi demand are as follows:
So last year, we spent a lot of engineering and management resources solving supply chain issues: rewriting code, changing our chips, reducing the number of chips we need, with chip drama central. And that was not the only supply chain issue. So there's just hundreds of things. And as a result, we were able to grow almost 90% while at least almost every other manufacturer contracted last year. So that's a good result.
Tesla is lean and willing to skip many validation cycles that other auto OEMs will not. Despite that…
And just to repeat Drew's point, we still expect to be partly or primarily chip limited this year. So that's the thing that's actually the driver. And that chip limitation should alleviate next year.
Cars don’t stop being chip limited in 2023. GM has said it, Ford has said it, Tesla has said it.
Continuing with the trend of non-semi companies that are doing semi, Apple. Apple kicks way more ass in semi of course, and their supply chain for semi is nuts. It would take teams of people to fully track it. I try my best. If you recall from yesterday’s post, we brought up some pretty gnarly inventory builds across the apple supply chain. Well their earnings here were great news. Apple crushed it.
The good: iPhone consensus was $67.82B and actuals were $71.63, so $3.81B more iPhones were sold then the street expected. This is a massive amount of semi when you go look at the bill of materials teardown for an iPhone. This also means the supply chain is healthy in terms of inventory levels. Mac’s beat street consensus by 14.3%, with $10.85B of sales the M1 Mac transition has led to meaningful share gains.
The bad: iPad did poorly with a 10.7% miss, but that isn’t all too shocking given rumors of Apple deprioritizing iPad for iPhone in the early and middle parts of last year.
The ugly: No guidance was given. Are supply chain issues gonna hammer Apple? They are masters of this, but one has to imagine at some point they get clobbered by them.
Speaking of supply chain issues…
We spoke about Lam Research supply chain issues on the last post, they had a specific culprit that cost them $200M+, which we named in the subscriber only section. For KLA, that supplier is not really in the picture. Despite this, they did deal with quite the crunch. The same story as autos applies here. Supply chain issues will plague the industry until next year, gradually getting better. Q1 revenue guidance is being held down 8% to 10% due to supply chain issues.
Before you go blaming semicap, think about it… You sell a bunch of multi-million dollar technical equipment that arranges atoms perfectly to make rocks think. It has to be kept perfectly unblemished and transported across international waters in planes. It then must be assembled and calibrated in someone else’s factory that may or may not be shut down due to Covid. Of course supply chains will explode when faced with the greatest supply chain crisis in at least a generation. This is exacerbated by the fact that the business is twice as large as it was pre-Covid.
Process control intensity seems to continue to increase in the long term. I do expect KLA to have a light amount of undergrowth versus Nova and Onto. KLA ranks growth markets by Foundry/Logic, 3D NAND, then DRAM. This is consistent with other capex announcements. In the timeline of a new process, metrology ramps first, so Intel will be spending an outsized amount on metrology. To be clear, their spend on metrology as a percentage of total spend will be higher than TSMC due to the larger number of node transitions.
We feel confident in our ability to grow throughout the year with total company revenue growth exceeding 20% and semiconductor process control systems revenue to outperform WFE growth again.
H2 is going be to be strong given how weak Q1 is due to shipping nightmares and Malasyia/China covid shut downs. I like KLA, but I think it is still a little too expensive compared to peers. It’s also not that expensive given the wafer fab equipment spend environment going forward... They are only trading at 17x consensus forward, and consensus for all semicap is a bit low. KLA has grown 7 years in a row now, so… they probably deserve to not be looped in with other players who are a bit more cyclical. KLA also has better margins. Services and software are a bigger part of metrology and inspection versus other types of semiconductor capital equipment.
But in a historical context, if you went up as much as our revenues have, we would see volume discounting happening with our suppliers, which would bring down some of the costs, and that's not really happening in this environment.
So there's a little bit of a cascading that happens whenever you have a disruption in a facility that's already running full out, I mean, equipment is running 24/7, people are working up to legal limits in terms of overtime.
This environment is nuts!
So if you think about some of the big new greenfield projects that have been announced, those are '23 projects, not '22. So I think the combination of those that are related to both supported from the regionalization efforts driven by things like the CHIPS Act in the U.S. should that come to pass. But even the other projects, which aren't dependent on that, are really much more about '23 than they are about '22. So as we look at it now, we do have pretty good visibility out through the end of the year and even into the first part of next year for the demand to continue to be very supportive of our business.
2023 will be a growth year!
And over the long run, we believe that WFE, given rising capital intensity will grow faster than semiconductor revenue in terms of a long run trend.
KLA saw a huge China deceleration from 33% to 23%. I blame lockdowns, because China sure isn’t stopping on equipment ordering.
Moving forward to MKS Instruments, the largest subcomponent/system supplier to semiconductor capital equipment firms. They did a good job executing in the 4th quarter, but Q1 we get more supply chain issues. They expect to be slightly down or flat compared to the 4th quarter in the semiconductor market.
We continue to execute on our strategy to gain share with key lithography, metrology and inspection customers. Our photonics sales, the Semiconductor Market exited 2021 at well over a $300 million annual run rate. We continue to progress on additional design win opportunities.
A big chink of revenue is coming straight from metrology and inspection tool providers. This unit grew 50% year over year organically, but with an acquisition, it grew 90% inorganically. I like MKSI, but I would simply caution do not go to crazy trying to understand every single segment, subsegment, and competitive/pricing dynamics within them. Trust me, you never will. It’s a deep dark rabbit hole.
On trailing edge semi supply:
It's still the kinds of electronic components, they're kind of from legacy fabs. So it's really not necessarily the most advanced types of electronic components. And these components are shared in multiple industries, and that's where we're seeing the constraints. So no change really from the past.
We missed this, but MACOM sold their stake in Ampere and recorded a nice profit. Very curious to know which party had the call option on the stake. They must be incredibly bullish. SemiAnalysis has heard some great things about the 5nm based Siryn which utilizes a custom developed core that can be deployed in huge core counts without sacrificing too much silicon area.
Next, I'd like to provide some additional information regarding the December 2021 sale of our less than 10% equity interest in Ampere Compute Holdings or $127.8 million. Our equity interest in Ampere was established back in 2017 when MACOM divested applied Microcircuit Corporation's Arm-based compute business. MACOM was a passive investor in Ampere, not involved in its operations, and we had limited rights.
One of Ampere's other limited liability members exercised its call option, which was established in 2017. The $127.8 million cash consideration amount was also established back in 2017. Our fiscal Q1 2022 financials reflect a $118.2 million GAAP gain recorded as other income. This gain represents the difference between the $127.8 million of proceeds and our approximate $9.6 million carrying value of the investment. There was substantially no tax impact as we were able to utilize net operating losses to offset the gain. This transaction further strengthens MACOM's financial position.
Silicon Motion Technology
SSD controllers have been tight for a year and half now, so this could be a knockout the park. At current pricing and increase in capacity, they can grow 20%-30%, but they believe they can grow much more if they secure more wafer supply from TSMC, raise price, and shift mix. They had a similar guidance last year and ended up growing 71%. This will not be repeated, but there is meaningful upside to Silicon Motion.
Client SSD market share now sits at 35% to 40%, sounds great, but long term I worry as Samsung and SK Hynix want to fully in source controller supply. Despite this, in PCIe Gen 4 SSDs (higher end and higher margin), Silicon Motion expects to be at 50% share by the end of this year. The real story to watch is the datacenter SSD story, especially PCIe Gen 5. ASPs and margins here are much better.
The wafer supply story is quite interesting. Silicon Motion is actually receiving less supply in Q1 sequentially, but in the 2nd half, they will receive a solid uptick. The firm pretty firmly says their bottleneck to growth is wafer supply, and this isn’t just short term, but the next 3 to 5 years. Wow. Talk about a long, protracted, tight market. Later on in the call, the founder/CEO backpeddled to say they can keep the trajectory if they can keep execution going, recruit talent in R&D, and get sufficient wafer supply.
We are in severe shortage in certain technology nodes, including 55-nanometer and all advanced technology nodes, 16-nanometer and 12-nanometer. However, I think we are comfortable in 28-nanometer wafer supply. We are just balancing [40/45-nanometer].
Frankly speaking, all the major new PCIe Gen4 are 12-nanometer technology node. And we just don't have enough wafer to support the demand to fill. If we have sufficient, I think we will revise guide immediately. For -- in addition is our UFS controller is 3.1 that uses TSMC 16-nanometer. We also have a severe shortage. Even TSMC increased in amount compared with the 2021, but they're not enough to meet our customer demand in 2022. So we'll continue to work with our customers, with TSMC together to gain additional wafer supply.
As I mentioned, 28-nanometer, we are a little comfortable. We could do a better wafer allocation in 28-nanometer to gain incremental sales revenue even without the additional allocation in advanced technology nodes.
This isn’t perfectly applicable to TSMC’s utilization or overbooked rates. From other channel checks, 28nm and 40/45nm are the tightest and 16/12nm is also fully booked, but not by as much. Wafer price increases confirm our view, but these comments surely put our view into question.
I think all the module makers, they are expecting NAND price decline continually. So they hesitate to really capture very large volume of NAND, so they don't need a more controller inventory. It looks like the channel inventory now also very low in China, and the activity has become very -- is very actively. So this is a very positive move, and we see the backlog moving up very strong even with mature technologies, particularly in the legacy node. And this is helpful for us to increase our confidence China market will recover strongly from late Q1 to late Q2.
I share my thoughts on the stock in the subscriber only section.
STM preannounced revenue already, so much of the financial release was a foregone conclusion. Their Apple business did really well. As did the automotive and industrial. The TI read through we mentioned yesterday was good on industrial, but not good on personal computing/communications. We could repeat a lot of what was said already, but its already pretty clear.
STM could do great with Apple later this year given the chassis redesign and rumored changes to the sensor suite.
The SiC business has a big question mark on it for me. I share my thoughts on this in the subscriber only section.
Behind the subscriber wall on this post are two items which were mentioned earlier and also a juicy short idea. We are going to be executing on it this morning and expect to hold it long term. It doesn’t have too much short interest, so it seems to be a great candidate. Now also seems to be a great time to raise some capital to buy for far better companies.
Sorry for leaving out WDC and Seagate, the movement between the pair is incredibly interesting, but we didn’t have time for it today. Please let us know if we missed any other earnings.
Keep reading with a 7-day free trial
Subscribe to SemiAnalysis to keep reading this post and get 7 days of free access to the full post archives.