SiC as a Dog (Part 2)
SiC as a Dog (Part 2)
Silicon Carbide is looking sickly yet again. This time it's not Tesla but end markets, particularly industrial.
In March, I wrote probably one of my favorite headlines for a Silicon Carbide piece. So, I decided that if a headline is nice, you can use it twice. I now present you SiC as a Dog Part 2. Below is SiC as a Dog, part one.
The sickness in Silicon Carbide started at On semi. Let’s look through the results and see the drivers, the big deal, and what we should do about it.
The market freaked out about Silicon Carbide, specifically On semiconductor, saying that automotive demand for the year will be weaker.
However, for the full year, a single automotive OEM's recent reduction in demand will impact our $1 billion target and we now expect to ship more than $800 million of silicon carbide in 2023, 4x last year's revenue. In '24, we expect the growth of our silicon carbide business to double the market growth.
This large OEM is implied to be Tesla. On won a second source contract from them, and their piling inventory and price corrections are impacting SiC revenue. Meanwhile, STM, the primary first source for Tesla, seems to have no problem and is on track for ~$1.2 billion in revenue for the year. They plan to guide up next year.
For us, next year would be, let's say, a step ahead, consistent with our objective to deliver $2 billion in 2025. And so we have the capacity installed. We have the supply chain secure. And we have the customer base and backlog consistent with this objective. And I will communicate, okay, at Q4 earnings and in January, the objective of silicon carbide revenue we would have next year. But be sure it will be a consistent step with the $2 billion objective of 2025.
This guide down in Q4 is mostly based on the weakness in EV markets that Tesla has to compete in, particularly in Europe. Inventory burn is going slower than expected for European OEMs.
Now, ON Semi believes that the hiccup in Silicon Carbide is just temporary. They believe they can still grow Silicon Carbide 2x the market, and Silicon will be flattish for the year. The market is freaking out on these results, rightfully so. So far, the first quarter has not been the end of weakness.
Look, I think from an automotive, I would say it's a broader stroke as far as the inventory comment I made specifically with Tier 1s in Europe. As far as the EV, yes, it is a single customer. I wouldn't say it is broad as far as in this immediate quarter. But we still expect it to grow in Q4. So it is not a -- I don't want to paint it as any decline or any issue in demand for EVs. EV demand is going to grow. It's going to grow in the fourth quarter and it's going to grow in 2024. It just didn't grow as much as we expected it to. And that's demand-driven, whether it's short-term demand or anything different, we'll have to wait until we get closer to '24.
Meanwhile, all this happens while Wolfspeed’s ramp is finally happening at Durham. They expect to ramp 3x QoQ for Durham, and there is some slight softness and new Chinese entrants. I think there’s a bit of a disconnect between reality and what is happening, and I want to point out what the big swing factors are and what it means for one supplier in particular.
China is Ramping Capacity
One of the primary concerns most market participants have is that China is ramping up SiC in a big way. This dynamic was discussed on the Wolfspeed call. I have been writing a bit about the coming insourcing wave for Chinese automotive as well.
The coming EV wave means that China will do its best to source all aspects of semiconductors, SiC included. But today, SiC in China is not at automotive grade.
So there's a lot of noise about different folks coming online, but the demand for our materials is very, very strong. And where -- and obviously, constant communication with our materials customers, many of them are looking for extensions and expansions and things like that. And then obviously, we just reported the largest capacity reservation deposit in the history of semiconductors with our deal with Renesas. The feedback we're getting from folks is that China is doing a lot of investing in silicon carbide. They're doing that in silicon as well. But the feedback we're also getting is that they're not automotive ready at 150, let alone 200. So I really don't anticipate a demand being below supply for any time in the future, really the next couple of years for sure and probably the next half of the decade.
But broadly, China is driving some of the incremental weakness. This is from STM.
Well, for Q4, I repeat, clearly, we have seen on the industrial market, especially in China -- Asia, China, that the order booking, okay, entering in the lead time window were not materializing at our expectation. And it has mainly impacted the general-purpose microcontroller. But this is for Q4. But now, okay, we have to acknowledge altogether that we went back to normal in terms of lead time and capacity utilization for this kind of device and for semiconductor industry except very few product lines like silicon carbide as well.
The problem is that I believe that Western OEMs continue to underweight the likelihood that China can ramp up SiC. Most participants underestimated China’s ability to ramp leading-edge nodes without EUV. SiC is different because every single tool is not banned from use. China will climb this expertise curve a lot quicker than Western companies appreciate.
I think there is a dual dynamic that makes substrate suppliers most at risk. For leading SiC device OEM makers, as long as there are solid partnerships and self-sourced substrates, they should be okay. I am thinking of the STMs and ON semis of the world partnering with Tesla. However, the incremental and longer-term SiC market will not be as big as Western device companies hope.
Given that they insource most of their substrate demand and then buy some substrate on secondary markets (WOLF), I think STM and ON will not be as impacted as the companies holding the huge capacity bases. WOLF will be the company that eventually gets caught in the capacity crossfire in the late 2025s. While they will likely be the leading quality SiC wafer then, the problem is that they do not make money, and they hope to make money in the late 2020s.
This dynamic will be dampened by the new capacity coming online. I think the hopeful FCF they can make from their ramps of JP and Durham will be short-lived if they keep reinvesting it in the long term. I think Chinese SiC is being underestimated. They will figure it out; making a SiC wafer is not as hard as making a 7nm device.
Solar is Strong at ON, Weak Generally
I want to talk a bit about the diverging aspects of SiC demand in solar. Industrial, particularly solar, helped ON accelerate SiC demand earlier this year. This continues for now.
Energy infrastructure remained healthy in Q3, driven by the continued adoption of solar and energy storage solutions. We remain on track to our full year projections with nearly 70% revenue growth over 2022 and we expect the growth to continue in '24 as demand for our hybrid modules with silicon and silicon carbide solutions for this high-growth industrial megatrend remains strong.
The fear is that while residential is weak and utility-scale solar continues to be strong, next year, demand tops out. There is a case where next year’s demand for solar is flattish at the utility scale, but this is a current debate happening in the solar stocks.
Now, this diverges slightly from WOLF, who called out weakness in industrial and China. If this broad-based industrial weakness spreads to solar, that is the next leg for On. I don’t think it will be that meaningful.
We remain confident that the demand from automotive customers will remain strong while we are seeing some softness in the industrial space, primarily in China and Asia.
But the real reason I want to write this note is that one supplier is finally left in the dust. It’s almost the perfect storm for a SiC supplier, and the shares, while weak, are probably still a bit too expensive from here. I have a lot more on this dynamic behind the paywall.