Microchip Technology Incorporated (MCHP) Earnings Call Transcript & Summary

All right. Good morning, everyone. My name is Toshiya Hari. I cover the semiconductor space at Goldman Sachs. Thank you all for joining us. Very honored and very pleased to have the team from Microchip with us this morning. We have Steve Sanghi, Executive Chair; and Eric Bjornholt, SVP and CFO of the company. I've got a bunch of questions, but we'd love to keep this as interactive as possible. So to the extent you have questions, please have them ready. First of all, Steve and Eric, thank you so much for joining us.

Thank you, Toshiya.

You're welcome.

So I wanted to start off with a near-term question. It's been a little bit more than a month since you reported results. You gave guidance, revenue up sequentially 5% for the September quarter. I think you spoke to visibility and sustainability into December as well over the past 5 months. A lot has happened on the macro front, and I'm sure the audience is thirsty and hungry for an update. So what do you see in the marketplace? Any deviations from what you guided to 30 days ago, 45 days ago?

So first, thank you for having us here. Since we had the last earnings call, the business environment hasn't really changed very much. We're pretty much seeing really what we were seeing before. The demand is very, very strong. Backlog is very, very strong. I think we said in our earnings call that in the June quarter, we had more business delinquent than the one we shipped, which really means over $2 billion of delinquent going out of June quarter. So such a strong backlog, and that thing will happen again going out of September quarter. So for the current quarter, we have really no turns to take. The backlog is solid. There's a huge amount of backlog that is delinquent. And if there's any dislocation, let's say, shutdown in a city in China, like it's happening in the last week in Chengdu in the western part of China, that product can easily be redirected to another customer because it's such a strong unsupported backlog. So the environment is very, very good. So as far as the current quarter is concerned, the current quarter revenue numbers are really in the bag. So now you look at the next quarter. The backlog is, again, extremely strong for next quarter. We'll have a couple of billion dollars of delinquent backlog rolling into the start of the next quarter. And next quarter, we remain capacity constrained. So the next quarter guidance is really much more dependent on what new capacity we can bring online and is really not as much dependent on demand. On the capacity front, we know our internal capacity is increasing, and we're getting some additional external capacity also as somewhat capacity is being freed up in the foundries from certain people where their businesses have gone negative. So combined effect of inside and outside capacity increase, we're fairly optimistic that next quarter, we are sequentially up again.

Got it. Steve, a follow-up on your last comment on the supply side of the equation. Again, you spoke to a constrained environment for the rest of '22, and I think into 2023 as well. In which areas of your business are you most constrained? Where are you seeing some of the constraints ease, if you will? And what are you doing internally as well as externally to improve supply going forward?

So I think on large majority of products across the company, our lead time is 52 weeks. We have a huge PSP backlog, which is well in excess of 50% of our total backlog. And on some of the key capacity corridors, just the PSP backlog is in excess of 100% of our capacity, which means there's no capacity left for anybody who was not on PSP. And we're not seeing any let up on that. We think the business remains demand constrained well into 2023. In fact, we can't forecast when we are not demand constrained in 2023.

Supply constrained.

I'm sorry, supply constrained in 2023.

Got it. On the PSP program, I'm sure you guys are getting a bunch of questions. We do as well. I was hoping you could take us back to Feb of last year when you circulated your letter to your customers about the PSP program, you introduced the program. What was the catalyst in introducing PSP? What did the customer respond so far? What has been -- what has it been like? And to what extent will you enforce these contracts if and when customers come to you and ask for cancellations or push-outs?

So we launched the PSP program in February of 2021. And the intention was that since we were supply constrained, we give customers the opportunity to give us a 1 year of noncancelable nonreschedulable backlog. And if the customer would be willing to do that, then we'll give them preferential supply, they'll become very good customers. They're giving us good visibility and we can acquire the raw material ahead of time, build the product and give them a very good service. And we thought there will be only a handful of large customers who are able to make a 1-year commitment and on a noncancelable non-reschedulable kind of commitment, but that's not what happened. We had thousands and thousands of customers direct as well as through distribution sign up for the PSP program, and PSP backlog rapidly became well in excess of 50% of our backlog. That's sort of not what was intended. So the program was successful beyond belief. Now it gives us a lot of benefit, when you have that much backlog, you can build a product in our factories in larger batches and don't have to change setups all the time. You can confidently acquire raw material, wafers, substrates, [ lead trains, ] material needed to complete the part and not be concerned about that part of the demand changing. So it has given us huge benefits in the factories in our efficiency and some of the very high gross margins really reflect that efficiency. So now where we are is the capacity increase we had forecasted 2 years ago, all of it didn't happen because the lead time for equipment went longer. So we had made PSP commitments to customers based on this additional capacity coming in. Additional capacity did come in, but not as much as we had intended. So therefore, there is a lot of PSP backlog that has gone delinquent. On many of the capacity corridors, that PSP backlog is well in excess of 100% of our capacity. So that's where we are today. And your question was in terms of cancellations, we are seeing negligible cancellations, request for cancellations out of PSP, and majority of them we're not honoring it. The only exception where we will honor a request to reschedule or cancel would be when we can take that product and save another customer who is in a desperate situation to get parts, essentially their line is down. In that case, we'll be making 2 customers happy, take it from one and save somebody's lines down. But otherwise, PSP backlog is not cancelable. Now remember, the argument is when we took the PSP backlog from a customer, we made a conscious decision to commit our supply to that customer and tell some other customers that we don't have any parts for them. So the customer who placed a PSP backlog has benefited now for nearly 1.5 years from having preferential supply from Microchip. And now somebody were to come and say, well, just take my cancellation of PSP, then it has no teeth in the program. And next time around, you could get the benefit of it, but have no obligation from it. So PSP order is not cancelable. We have no plans to take any kind of broad-based cancellation or pushouts from PSP.

Interesting. And Steve, as we were discussing last night as another benefit to PSP, it adds an element of discipline on the part of your customers when they place orders, right? Because these are bigger batches, if you will.

So you talk eloquently about how PSP backlog is higher quality, maybe you can talk about that.

Sure. So I mean if we had a standard cancellation window for all orders, let's say, it's 30 days or 90 days, we would have a mountain of backlog that would be stacked up right beyond that cancellation window and customers could just continually push that out, cancel the orders, do whatever they want with it with no skin in the game. So it gives us a significant advantage in terms of knowing that customers have a commitment to us, putting capital in place, putting the people in place, adding the raw materials to service them. And the customers, in most cases, will need to go to a higher level within their organization to get a commitment of that type, right? If Microchip was making a commitment for 12 months out in time, myself or another senior leader in the company is going to be signing off on that. So we think the backlog that we have is significantly better than what it would be without PSP. Backlog could even be higher if we didn't have the PSP program, but we wouldn't know what was real and wasn't real. And every month that goes by, customers in the PSP program are required to put their next month of orders on with us. And so we have a good leading indicator also in terms of what's happening in the marketplace. And if we see a customer or a group of customers is going from buying at a certain rate, and that is being reduced or increased, we can make the necessary adjustments in our business from an investment perspective to respond to that.

Got it. Many of us who have been following the semiconductor industry have been trained to expect pricing to be down low singles, mid-singles year-over-year. Obviously, that hasn't been the case in the past 2 years. And the big debate that we have with investors is, is it different this time? Or do we go back to pricing being down in year 2, 3 and 4. How are you guys thinking about the sustainability of pricing. And to the extent you do think it's different this time, why would that be the case?

So I think the price increase has largely been a result of input cost increasing and us then passing those input costs to our customers, margined up but essentially passing the input cost. We don't see input costs coming down. We don't see people will take lower salaries next year or year after. We don't see cost of equipment going down. We don't see cost of other gasses, chemicals, raw materials, industrial materials costs coming down. So with the input costs really staying high and they're still going higher, just yesterday's inflation report, inflation was 8.3%. And I think, as we speak, industry is actually talking about another round of price increases right now. So I don't really see that these prices come down because input costs are staying higher. Now secondly, about 12 years ago, 10 years prior to COVID, at Microchip, we begin to talk to our customers that there is no longer year-over-year price decrease. We are the only industry and it's crazy industry that works hard and just give the lower price every year. And we stopped doing a year-over-year cost reduction, especially on all the older products. On some brand-new product where the volume is very low, as the volume rises, we'll give a small price decrease but all the older products, we would not give any price decrease. So 12 years ago, it will work maybe 20% of the time. And 5 years ago, it was working 90% of the time. So well prior to COVID, we had established a relationship with our customers that we won't give you a year-over-year price decrease, but we won't gouge it either. If the capacity becomes short, we won't gouge it either. And that's the relationship we had on proprietary products. So they can design with confidence that our price would be thoughtful and our price would be fair, and that's what it has been for a long time. Now enter COVID, nobody had expected all the input costs will go up. And as we explained to the customers how we were passing the input cost increase, customers understood it: they were seeing it in their own business. And most price discussions really have not been that difficult. So as we go forward, we believe we will simply return that to our old way, that there is no year-over-year price decrease. But there's no increase either if the costs are constant.

And Steve, based on your current conversations with your foundry suppliers, is the expectation for input costs or wafer cost to go up again in '23? Or is that still TBD?

It is TBD, but the inclination seems to be and at least 1 foundry is out and open that there's another price increase in 2023. And I would think that, that would result in another price increase passed on by the semiconductor industry to its customers.

Got it. Let's talk a little bit about your long-term growth drivers. You hosted an Analyst Day last year, you talked about a 6% to 8% CAGR. You bumped that up in March, I believe, to 10% to 15%. You've been pretty vocal about these mega trends. I was hoping you could kind of level set the audience by going through some of those mega trends and perhaps rank order some of them in terms of your expectations.

Okay. Well, I'll take the piece of it on what we did to update the model. So we had taken at our Analyst Day back in November and said, "Hey, we think that we can grow at twice the industry growth rate." Industry growth rate historically, if you look back at the last 10 or 15 years, it was 3% or 4%, and that's where we came up with those numbers. When we looked at using fiscal '21 as a baseline, we reassessed that and said, "Hey, for the next 5 years through fiscal '26, we believe that we can grow at 10% to 15%. Fiscal '22 was a great year for us; fiscal '23, which we're in now, has been an excellent year. And so we're off to a really good start with that. At the low end of that range, we factored in a recessionary period, and we're tracking well to within that range. I'll have Steve talk to the mega trends.

So at the Analyst Day last year, we introduced how we were doing across 6 megatrends of the industry. And those were 5G, IoT, data centers, electric vehicle, artificial intelligence and machine learning. And the final one was advanced driver assist someday leading to the self-driving cars. So those were the 6 megatrends, and we identified that our design wins in those 6 megatrends were at about twice the rate of the rest of our business. So we were really penetrating hard into those megatrends. And the revenue from megatrends was about 1/3 of our total revenue. So significant revenue was coming from megatrends and that part was growing at about twice the rate. So just recently, we have updated the megatrends. We have taken one mega trend, which was artificial intelligence, and what we are finding is it is a mega trend but it is seen in advanced driver assist. It is being seen in IoT. It is being seen in data centers. So all these other applications are taking advantage of artificial intelligence. And therefore, rather than calling artificial intelligence as a stand-alone megatrend, those designs with artificial intelligence are actually being counted in other megatrends. So we have removed artificial intelligence as one of the megatrends and have replaced it by sustainability. Sustainability of power generation, water, sustainable planet, green energy, emission reduction and all that, all that equipment uses a large amount of Microchip products. So that's sort of a new megatrend we have introduced to the market just in the last couple of weeks, I think Ganesh did that in another conference. So that's what's happening on the megatrend. I think we're doing very, very well.

Got it. Steve, you guys have been quite acquisitive over the past 5, 10 years, and you've transformed the company from a microcontroller centric company to more of a total system provider. Can you talk a little bit about that evolution and how you're differentiated competitively now that you have all these pieces, if you will, to the broader picture?

So we began the acquisition trend back after 2009 when we were a $740 million company in that year, making microcontrollers and a handful of small analog parts. And we were competing with companies substantially larger than us -- back then, Linear, Maxim were independent companies, Analog Devices, Renesas, STMicro, Infineon, NXP and others. And I was CEO back then, and I found 2 issues. One, we were subscale to these companies, and we had a scale disadvantage. And second issue was we will win the microcontroller in a customer board but then we're surrounded by all the competitors who are providing all the analog, power management and everything else. And some of these competitors also make the microcontrollers. So when they visit a customer talking about the other parts, there would be knocking on our microcontroller. So the strategy we deployed at that time was, number one, in the subsequent decade, we need to scale the company 10x. And number 2, in the process, build out everything that goes around the microcontroller so we can elbow out the competition. And when we are there, we are the only one there, and there's nobody else -- at least try to get there, have a majority of the pieces only come from Microchip. So in the subsequent decade -- and that's what we achieved. We took the company from $740 million to over $7 billion over 11, 12 years. And in the process, we either built or bought, acquired companies that provided all the sensors to op amps, to converters, to power management, to connectivity, USB, Ethernet, WiFi, LoRa, Bluetooth, memory, flash, SRAM. So today, we do everything. And today, if you look at a customer's board on their existing design, if we have 2 parts from Microchip, in their next design, we have 12 parts from Microchip. I mean, it's just dramatic. And that's what we call total system solution. So when you talk about the differentiators, I believe our #1 differentiator is rapidly becoming the TSS. And we have organized to sell it in a way that the customer cannot break that bundle. If they break that bundle, some of the benefits go away, cost of benefits as well as warranty and others because much of our software is porting into it and they can't use it if they break the bundle. So TSS is a differentiator. Our focus on megatrends is a differentiator. And those are -- and we have one of the broadest product lines today in microcontroller and other things. So those are the hard differentiators. And then you have some soft differentiators like culture and people, which are the backbone of the business. But everybody would say they have great people. So I think that's a soft differentiator. But our culture has been largely talked about over years in the industry -- and we have awards in roughly every major geography we operate in as one of the best companies operating culturally with people through surveys and all that. So I would say those are the differentiators.

Steve, the other question that we often get more recently, given the decline in valuations is could Microchip revert to buying companies again. It sounds like you're very comfortable with where you are in terms of all the pieces that you have. But is the picture complete, and it's more about organic growth going forward? Or could you contemplate M&A?

Yes. So M&A had a purpose, like I described, to complete the solution, and we have completed the solution. There are no gaping holes in our portfolio today. We continue to just design in deeper and more and more products and higher and higher performance and lower and lower power in each category of products. So M&A is really not -- not really what we're pursuing. If we do any M&A, it would be very small tuck-in where we go buy a 20-people design teams somewhere with a specific expertise either in a data center or FPGA or some other unique discipline. But there are no plans for a large-scale M&A like we did in the last 10 years.

Shifting gears a little bit. TI has been pretty vocal about growing capacity in the second half of this year as we speak and into 2023. One fairly common question that we get is, could that be a risk not just to Microchip, but for the broader MCU and analog industry from a supply-demand perspective. I know it's not as simple as that, and you don't perfectly overlap, but how should investors think about that dynamic into 2023?

TI has been a very disciplined -- gross margin discipline company and very well-run company for a long period of time, and they are a very good competitor. And I don't really expect that to change. There have been players in our industry that were totally price undisciplined. And we bought many of them and subsequently adjusted the prices and successfully integrated them into Microchip. Over years of having run the company, I never looked at TI as a real price aggressor or price indisciplined. So I don't think that's likely to change. So I'm not really worried about TI.

Yes. I think maybe your question is a little bit more capacity in nature, and they have a different business model than us. I think they do 80% to 85% of their fab in-house. We do about 40%. And so it's just a different model. We think we are making the investments in capacity that are appropriate for our business to drive our growth.

Got it. And then a couple of financial questions, maybe more for you, Eric. But on the gross margin side of things, you guys have executed really, really well. Guidance for the current quarter, I believe, is 67.5% at the midpoint, which is at the low end of your long-term model. So you're pretty close to being at a steady state, if that's the right way of thinking about it. What are the potential puts and takes that we should be aware of, we should be thinking about going forward? I know there's a lot that goes into one number, but what are you guys focused on? What are some of the initiatives inside the company?

Yes. So I mean the things we laid out at our Analyst Day last November really haven't changed. We're continuing to invest in capacity. We have intentions of bringing the 40% that we do internally for wafer fab up to about 45% over time. On assembly and test, we do roughly 60% of that manufacturing activity in-house today. We expect assembly to move to about 70% and final test to about 80%. So these things are items that we think incrementally move gross margin as we go forward. Steve talked about the PSP program earlier and how it makes us so efficient today in terms of knowing what to run on the factory, every part that we're running is being shipped to a customer. And that will change over the course of cycles, but we're being efficient today. We've got to the low end of our long-term model quicker than ourselves and investors thought we would. And we'll continue to make improvements from here.

Got it. And then this is more of a hypothetical question. And based on the comments that you've made so far today, it sounds like your base case internally for '23 is more of a soft landing. But to the extent things get a lot worse and your revenue is down 10%, plus or minus, in '23 versus '22? How should we think about that gross margin profile changing into next year?

Well, I would disagree. I don't know how we gave the impression that our base case for 2023 is soft landing. Our base case for 2023 is growth. It's not soft landing. What we have said is we see the macro deteriorating, many companies announcing interest rates rising and things happening in the economy. If it were to be the case that the macro finally catches up to us, then what are we doing to make sure the business soft lands? So that soft landing scenario for us is a what if. But our current plans for 2023 is actually growth. We see solid backlog all through 2023. There are many, many products we don't even catch up on demand supply -- don't even catch up through even the end of 2023. So I just wanted to make that correction.

Yes. And we've created some historical slides that are on our website to look at the last 15 years of gross and operating margins through cycles. I think you'll see on there on the gross margin side that we've experienced like a 300 basis point decline in gross margin. Times were different back then. We were doing acquisitions during that full time period where we'd acquire companies with lower gross margins, so that impacted some of those dips. And we had a larger amount of internal capacity that we're relying on. So when things fell down from a revenue perspective, we had larger underutilization charges as a percentage of our total cost of sales. So the company has changed a bit over time, but we think our gross and operating margins stay high throughout the cycle. It doesn't mean that there'll be -- won't be ups and downs, but we think we'll be able to moderate that to some extent.

Appreciate that. Very helpful. We have a little over 10 minutes left. I wanted to pause here and see if anyone had any questions in the audience. If you do, please raise your hand. I'll keep going. I guess the CHIPS Act is another topic that comes up in conversations, and we discussed this last night a little bit as well. Can you talk about some of the benefits there and then the Inflation Reduction Act, I guess, is the potential offset. But maybe, Eric, if you can talk about the pluses and minuses there?

Maybe I'll take the tax side of it, and Steve might want to comment on other elements of the CHIPS Act beyond the investment tax credit. So CHIPS Act is favorable to Microchip on the tax side, right? We get an investment tax credit of 25% of the dollars that are invested in U.S.-based semiconductor facilities and equipment. The net benefit turns out to be a little bit less than 20% because you have to reduce the amount of your depreciable base within those assets for tax purposes by the amount of the credit. But anyway, that's just mechanics. Overall, in periods where we are investing a lot in capital like we are today, this is going to be quite beneficial to us. There's this other act called the Inflation Reduction Act that introduced an alternative minimum tax of 15%. That is negative to Microchip, but we view the combination of the two to be modestly favorable to us in times of capital expansion and modestly dilutive to the tax rate or negative to the tax rate when we're not spending as much on capital. But overall, it's manageable for us, and we're happy that CHIPS Act is bringing this ITC and that accelerated depreciation for tax purposes remained in place. So do you want to talk a little bit about CHIPS Act from a funding standpoint?

Yes. So CHIPS Act has allocated $52 billion for domestic manufacturing -- to subsidize domestic manufacturing, so people build their fabs in U.S. rather than building it elsewhere. The rules for applications and all that haven't come out yet. We're hearing it will take about 6 months for the procedures and rules and all that to come out, and they're trying to assign people in the government who will manage this process. I expect it to move at the rate -- at the pace of the government, which is often very slow. But at the end of the day, if you make investments in U.S., then you would be able to get a 1/3 subsidy from the government. So if you spend $1 billion in building a plant, you get 1/3 of it back from the government. And the piece that -- 2/3 of the piece that you spent, I believe you also get an investment tax credit on that, which is another 25%. Although rules on that are not totally clear, there is some restriction on double dipping, but the rules are not clear. So I think overall, it's really good for the semiconductor industry. China, Japan and Europe have historically helped their industries with a large amount of subsidies. Such subsidiaries have never really been available in the U.S., and it's the very first time. And I'm a free marketeer. I'd rather have no subsidies anywhere in the world. And that would be the best. But if China is going to do it, Japan is going to do it, Europe is going to do it, then the CHIPS Act is good for U.S. So that's where we are.

And we've spent a lot of time over the course of the last year educating the Department of Commerce on the importance of the entire semiconductor supply chain, and I'll refer to this as more trailing edge technology and how a system can't be built without it. The other thing that we've highlighted is that Microchip is the largest supplier of semiconductors to the U.S. defense market. And so we think there is a vested interest that the U.S. government has with Microchip. And so we're going to apply for as much money as we can get and see what happens.

And can you guys remind us what you're thinking on CapEx going forward, capital capacity expansion plans, both in terms of wafer processing as well as assembly and test? And does the CHIPS Act in any way change the way what you're thinking about those things?

So I'll try to address your question. You can expand on it if you want me to say anything else. But we're investing between $500 million and $550 million in fiscal '23 in CapEx. We had a large CapEx last year, too, which was a little bit more back-end manufacturing weighted. Lead times on equipment, installation time is less for the back end. So it's a little bit more front-end wafer fab that we're focusing on this year. I talked about how we're trying to go from about 40% to 45% for internal manufacturing on wafer fab. We're making good progress. We're expanding our clean room space in Oregon. All of our factories are increasing capacity as we speak, and that's allowing us to produce higher shipments each quarter and give us the confidence to even guide that the December quarter is going to be up. What else do you want to talk about on capacity front?

The CHIPS Act, that being passed, did in any way change your thinking?

CHIPS Act today is not in the thinking in guidance we have provided to the Street. So this $550 million we're spending is largely in our existing factories and our 3 fabs and 3 assembly and test plants. CHIPS Act rules are not all clear yet and it's only 2 or 3 weeks old. And like I said, it will take about 6 months before the rules come out. So any CHIPS Act funding is really not in the next year -- between now and next year. And we're starting to think through internally with the CHIPS Act subsidy, what makes sense for us to build. And not quite ready to talk about it publicly.

Yes. So we haven't guided outside of fiscal '23, which ends in March. Obviously, CHIPS Act and the investment tax credit does become effective for us in the first calendar quarter of '23, so the last quarter of our fiscal year, and will have a positive impact on that. And hopefully, we learn a lot more over the coming months and can build our fiscal '24 plan for capital and share that with you with the inclusion of anything that's happened because of the CHIPS Act.

And new fabs, new capacity takes a long time to build, several years, and then ramp them. But the possibility could be that our internal capacity moving north of the 45% you talked about longer term, several years out. If we were to take advantage of it and do something with it, and we don't know yet, but we will look.

Steve, I guess, many of us have looked to you as an industry leader that has seen many, many cycles. You've called many cycles, you say it as you see it, if you will. I think many of us are perplexed, right, with the disconnect between some of the macro headlines and what you're saying and how you just described 2023. Collectively as a group, what do you think we're missing or underappreciating about the overall semiconductor market, more specifically, how you're positioned, PSP, TSS, those things make a ton of sense -- but anything else that we're perhaps missing?

I think what most people seem to be missing or not connect the dots is the difference between leading edge capacity and trailing edge capacity. If you go back a decade ago or more than that, foundries will build a leading-edge fab and give its capacity to the leading-edge companies like Qualcomm and AMDs and others, NVIDIA and others and fully depreciate it in 4 years. And then the leading-edge guys move to the next fab, which is the next technology and this fully depreciated fab would be repurposed for building trailing edge products, microcontrollers, analog, mixed signal and others. I don't recall a foundry ever building a trailing edge fab to build trailing edge product. Trailing edge capacity always came from leading edge guys moving to the next technology. With such a large shortage and the growth we have seen in the last couple of years, that math is broken and it's probably broken forever. Two things have happened. Number one, if you have to build a trailing edge fab today, its cost is 40% higher because of the depreciation; the older fabs had no depreciation. So unless the prices were to go up 40%, they've gone up some halfway, but they haven't gone up 40%, foundries are still saying they will not build a new trailing edge fab. So that's one problem. Second problem is technology has gone in different direction. The trailing fab are largely 8-inch. They have aluminum back end. The leading-edge technologies are copper back end. So a leading-edge fab becoming available, you cannot build trailing edge products because the metalization is not common, and there are other issues. Intel's 22-nanometer and up fabs are all FDSOI, they're not planarized technologies, and I won't get into the details. So the technology has diverged. So therefore, there is a trailing edge capacity and they're a leading edge capacity. And just to say, take a 180-nanometer chip and go build it on 22-nanometer technology, anybody who talks like that doesn't know what they're talking about. So I think with that divergence in leading edge versus trailing edge technology, all the capacity is being added on the leading and foundries are adding no capacity on the trailing edge. So therefore, IDMs like us and TI and NXP and STMicro and others, have taken on the task to add some trailing edge capacity ourselves. But we're not adding enough and we can't add enough. We only do 40% of our business, 60% comes from foundry. So if they won't add their [ fees, ] we'll make up some, but it's possible that the trailing edge capacity is constrained in our industry for a long time until this math reconnect, either prices have gone up enough or some other technology breakthrough where this math works. So I think when somebody on the leading edge guides down or has a preannouncement or something, the whole baby is thrown out with the bathwater. We're not seeing anything in our business that NVIDIA is seeing. We're seeing nothing in our business that some of the other companies are seeing. Yet that link needs to be broken across the leading edge and trailing edge lines. And I think I have talked about it significantly, I talked about it last night. But I don't think there are enough -- there's enough time being spent to educate the market, educate the investors about that difference. That is different this time.

Appreciate that. Very insightful. Steve, Eric, with that, I think we're officially out of time. Thank you so much for joining us this morning.

Thank you.

Thank you all.