Microchip Technology Incorporated (MCHP) Earnings Call Transcript & Summary

Greetings, and thank you for attending JPMorgan's 49th Annual Technology, Media and Communications Conference. My name is Harlan Sur, semiconductor capital equipment analyst for the firm. Very pleased to have the team from Microchip here with us today. Steve Sanghi, Executive Chairman; as well as Eric Bjornholt, Chief Financial Officer. It's been almost a month since the team reported record revenues, record margins and EBITDA, and it's been a busy earnings season. So I've asked Steve to start us off with some opening commentary, and then we'll go ahead and kick off the Q&A. So gentlemen, thank you for joining us today. And Steve, let me go ahead and turn it over to you.

Thank you, Harlan. Welcome, everyone. Before we begin today, I wish to remind you that in today's discussion, we'll be making some projections and other forward-looking statements regarding the future financial performance of Microchip. These involve predictions and the actual results may differ materially. I refer you to Microchip's filings with the SEC regarding some important risk factors about the company. With that, let me start with a few brief statements about our business. The March 2021 quarter that we just reported was a record in many respects, including revenue, non-GAAP gross margins, operating margins and EPS. We have guided the June 2021 quarter to even higher levels as we continue to execute our long-range plans, focusing on providing total system solution to our customers. Business conditions continue to be exceptionally strong, backlog remains at record levels for product to be shipped over multiple quarters, demand increases continue to outpace the capacity improvement we're able to make, and hence the gap between supply and demand continues to widen. With the strong demand and low level of inventory book at Microchip and at our distribution channel partners, we are experiencing constraints in all of our internal and external factories and their related manufacturing supply chains. We are continuing to ramp production in all our internal factories as well as investing in capital additions to expand our internal capacity based on the strength of our backlog, especially our noncancelable preferred supply program that we call PSP. We are also working closely with our supply chain partners who provide wafer foundry, assembly test and materials to secure capacity. Through the combination of internal and external capacity additions we have taken, we expect our overall capacity will continue to grow every quarter in calendar 2021 and beyond. While our capacity will continue to grow every quarter, we also believe that wafer fab as well as assembly test constraints are here to stay with us through calendar 2021 and possibly well into 2022. About 4 months ago, we launched our Preferred Supply Program, or PSP, to provide customers with supply priority beginning 6 months after their order in exchange for at least 12 months of noncancelable and non-reschedulable backlog. Customer response to the program has exceeded our expectations with direct customers and distributors alike. At the time of the earnings call, about 44% of our backlog was in the PSP category. Since then, PSP backlog has increased and is now about 50% of our total backlog now although it's almost 100% of our backlog in some of the most constrained capacity corridors. Additional PSP backlog continues to come in every week. This gives us a solid foundation to enable us to prudently acquire constrained raw materials, invest in expanding factory capacity and hire employees to support our factory rents. Finally, our Board of Directors is systematically moving towards a higher shareholder returns model, and we are rapidly deleveraging our balance sheet with a laser focus on becoming an investment grade-rated company. This concludes my update. Let me pass it back to Harlan for Q&A.

Thanks for the update, Steve. So I remember back at the earnings call, you said that in your 40 years, you haven't seen a time when the imbalance between supply and demand has been this significant. Ganesh talked about the fact that your view was demand trends were about 40% above your ability to supply at that point in time. You talked today about that gap actually continuing to widening. And I'm just wondering, the question for you is there are several things that I can maybe think about in terms of that gap widening. First of all, maybe if you could quantify, has it moved from 40% to 45% or 40% to 50%? And is it more a function of demand just keeps getting stronger and the unfulfilled backlog just keeps layering on as we move forward into the second half of the year? Or is it maybe more the supply side where your capacity increases are a bit slower? If you could just help us understand the dynamics there.

So quantifying the gap, at the earnings call, we said that if we had infinite capacity, we could ship 40% more this quarter. That gap has now widened to 50% and is continuing to rise. It's not stopping. I think it's going to go up further. And the gap largely has been -- the demand order rate has been stronger than the increases in capacity. As I mentioned, our capacities continue to increase, but demand is layering in faster than that. So there is -- on many products, there is no product available for the next 9 months. It all has been committed already. So for a new order coming, if the customer wants that part in June, we're scheduling it in April next year.

Okay. So it's more a function of demand. And you've talked about before, seeing strong demand trends in industrial, auto, consumer white goods, but we know that as we move into the second half of the year, for example, we've had Intel. We've had AMD and a few of our other companies, but cloud spending is starting to pick up as we move into the second half of the year. Enterprise demand is starting to pick up. We have the seasonally stronger second half of the year. So is it some of these other demand dynamics that are starting to sort of fill into the second half, that is you think probably the cause for the widening of the gap? Or is it just auto and industrial just continues to be getting stronger?

Both are correct. The broader the demand gap is continuing to widen. And data center, for one, is particularly going to see a stronger second half. We're seeing significant growth and backlog layer in. That was one of the segments that was digesting Huawei, and we didn't have shortages on the data center products in the prior quarters because we've been seeing that sector has been weak. And now that demand is strengthening and the backlog is very strong, and we are starting to see the signs of where we are going constrained even on that. So data center definitely is one of the driver and it's second largest section of our business. I think it's about 18% of our business. So it's meaningful, but it's not all driven by data centers. All the other sectors are continuing to see strength.

Okay. Perfect. And then on the PSP program, good to see the success and growing success there, right? As you mentioned, at the time of earnings, it was 44% of backlog. It looks like it's increased to 50% of backlog. I think as of early March, I believe you had 1,200 customers committed to that program. I don't know if you can update us, given that the backlog on PSP has increased so much. How many customers do you now have committed to the PSP program?

So we don't have a number of customers. There are thousands and thousands of customers that come to us through distribution. And every one of our distributors worldwide has put various devices and their end customers on PSP. The number is likely to be in thousands, and we're really no longer tracking it. We're simply tracking it by the percentage of backlog that's coming on PSP. When we gave you a number, we were tracking larger customers, smaller customers are very hard to track. But the PSP backlog today is about 50% of total. And given that just since the earnings call, it moved from 44% to 50%. You could do the math. That means very large portion of incremental bookings are coming on PSP. And even 50%, I believe does not include a large amount of orders that are sitting in that our planning group is overwhelmed and has not been able to process. Remember, we were not structured to process 3x the orders in a quarter. You have certain capacity of people and equipment and other to be able to schedule the orders. When the orders go in where the entire year worth of backlog is coming in, in 1 to 2 quarters, then our planning group, scheduling group is just overwhelmed. And there are a large number of orders sitting there that need to be scheduled and whole bunch of those are PSP, so the gap is likely to widen.

So given the success of the PSP program, given -- and I'll ask you about lead times in a second. But given what I think are extending lead times, the team has very good visibility into the second half of this year and -- I mean, and obviously, good visibility into next year. And so if you layer upon -- on top of that, your capacity expansion outlook and then you take into account there's going to be a period of time where customers are going to have to replenish inventories. I mean when do you see the supply/demand gap closing? I mean is it first half of next year? Is it middle of next year? Or is it going to take us through all of calendar year 2022? Would love to get your views on that.

So I think in the November conference call -- November 2020 conference call, I first mentioned about significant constraints developing and said that I don't expect those constraints to go away in 2021, stay at least through the end of 2021. A lot of people even questioned that call and said how could we see that a year ahead. And I was seeing that kind of strength. So now we are 6 months from that call. And I think I'm moving the time by at least that 6 months, which means I think we'll be fighting these challenges, at least through the start of the second half of 2022, and possibly go even further into late part of the 2022.

I appreciate that. And can you just give us a sense on -- I know your lead times vary quite a bit, so if you want to give us a range. But can you just tell us how much average lead times have expanded maybe year-to-date? And do you expect them to continue to rise on a go-forward basis?

So I think -- I've always found that lead time, you can't give a number, which is an average. There'll be like, I take your 2 feet and put one in a 0-degree water and other in 150-degree, average is 75, you should be very comfortable and you will be not. The lead times are similar. There are products which you can buy off the shelf. I have lead times of less than 4 weeks. And the other is that I can't make it in 50 weeks, and it's all over the place. We make hundreds of thousands of SKUs. And lead times are all over the place. It largely depends on a given product. If the backlog is strong, then every new order coming in is going to be scheduled outer and outer in time. So a lot of the extremely constrained corridors have lead times of 40 weeks, 50 weeks and higher. And a lot of corridors, which are not as constrained, the lead times are shorter, 4 weeks or even less.

On the supply side, the wafer front-end equipment suppliers we cover: Applied, Lam, KLA, for example, they've all talked about their extended lead times in leading-edge and in mature technologies. The Microchip team has talked about extending lead times and even some order push-outs for wafer equipment in the last call. Are you guys still experiencing equipment push-outs for wafer fab equipment? And if you place an order today for wafer fab equipment, is it fair to assume that you guys probably aren't going to get delivery of those tools until next year?

So I think it's all over the place. We haven't seen any additional push-outs since the earnings call if that's your question.

Yes. Yes.

We also buy a lot of used equipment because a lot of our internal production is on legacy technologies. So everything we're buying from some of the companies you mentioned, I don't know what's happening at 5 micron, 7 micron, 10 micron or even 22 or 28 or 40. Internal production is all 8-inch. And it's legacy technologies. So the equipment lead time, number one, are slightly shorter than some of the very advanced equipment, maybe. But we're not -- we haven't seen additional push-outs. But the lead times that are in the 8 to 9 months range, yes. So if you order something today, trying to acquire something today, you'll probably be getting it by March. But we have everything we need to grow our capacity between now and March. We already have it on order and schedule. And if it gets further pushes out, we'll see some issues, but I think they're not really -- we're not seeing further push out. And we are continuing to order additional stuff. As we speak, we're analyzing what we need in fiscal 2023 and what should we order that will arrive on April 1 onwards.

Perfect. We have a question from an investor here. How does Steve envision customer inventory management to evolve as the supply and demand comes back into balance? Will there be a permanent change to customer inventory management in the outer years?

So I think as we have said, the difference between supply and demand has been the widest this time. I do not remember prior instances, prior cycles where $1 or $2 of microcontroller stopped a $50,000 car from shipping. This is really the worst. Last time we saw that kind of impact was after the Tsunami in Japan. And Toyotas of the world learned their lesson from that and decided to really, number one, pay more attention to semiconductors and build a inventory safety stock for these kind of disturbances. And in the current cycle, those customers have done better than anybody else. I think U.S. and European automobile makers have shut downs this time around when the Toyotas of the world probably haven't. So this lesson would be learned by the rest of the industry this time, not only automotive, also industrial and appliance makers and the rest of all the supply chain up and down. So I think some of those lessons will be long-lasting. They may forget that in 5 years, but I don't think they're ready to forget that in 2022, '23 or '24. So I think this one will really raise the importance of semiconductors and keeping some safety stock for everybody, which will help prolong the cycle, which will help also soft-land it, because the distribution inventory is very low. Our internal inventory is very low. The supply chain, there's no inventory at the subcontractor. There's no inventory at the customer. And if all that has to be built up and made healthy, if excess demand were to stop today, it would take us several quarters just to bring microchips or [indiscernible] and finished goods and all that into a healthy place. It will take an additional longer time to take distributor inventory, which have fallen off the bottom, I think the 15-year low was about 28 or 29 days. Last quarter, they were 22. And it could go further lower this quarter because everything we ship is going out the door. So I think some of those lessons will take much longer to forget.

Another question from an investor. Steve, what's your take on double ordering trends or maybe padding of the orders? And is PSP really, in your mind, a deterrent to that?

So I think what we have said repeatedly is that double ordering or so-called double ordering happens 3 different ways. One way it happens is a customer places the order from 2 different suppliers. Now that you can't do in microcontroller because they're all custom. You can't take our part out and put a Renesas part in. Either you designed with us or you design with them, and nobody has the resources to do double design because of all the engineering shortages. So that kind of double ordering doesn't happen on 98% or so of our business, which is all proprietary products. The second kind of double ordering happens where a customer places the order from 2 distributors, placed the same order at Arrow and Avnet, let's say. We have systematically eliminated that by asking each distributors to register their design. So we ship the part to distributor at an elevated cost and then we give them ship and debit to allow their cost to go lower so that they can make margin in the shipment to the end customer. So if a nonregistered distributor tries to ship the parts to a customer, they would be shipping off a very elevated cost and make no money, hence making it commercially not viable. So we eliminate kind of that kind of double ordering. And then the only one left is somebody wants 100,000 parts and they order 110,000, I don't know whether the demand is 100,000 or 110,000. So that one we can't eliminate. So double, triple ordering, we eliminate, but slight excess, we don't eliminate. That's why I'm saying that currently, we are 50% constrained. And if 20%, 30% of it, froth were to go away from the top, we still have 20% more demand than we can make. And therefore, we're really not as constrained. So while there could be -- there's no double ordering in our business, while there could be some excess ordering in our business. One thing I'm sure is there is no excess fulfillment, and there's no double fulfillment because everybody is screening for parts. And there's no double fulfillment. And there's no capacity to double fulfill it probably for the next year, at least.

So let's move over to talk about products. The product leadership and total system solution strategy. So in your microcontroller business, record revenues in the March quarter. In calendar '20, you gained share again in all of your MCU segments, strong #3 global share leader. You've had a consistent track record of annual share gains and also articulated particularly strong gains in 32-bit MCUs last year. You've got your PIC architecture, you have your ARM-based architecture. I guess the first question is, any differences that you're seeing in market adoption between the 2 different architectures? And then in your general sort of overall MCU business, is the team confident on further share gains this year? And what areas are you expecting to potentially gain share in?

So, I think, Harlan, we have been consistent for a decade-plus that we don't sell a microcontroller based on architecture they're built on. The architecture engine is a very, very small piece of the entire microcontroller. Just like in a car, you don't really know what engine it has when you buy a car. You buy it based on features and performance and brand name. And you want a Tesla, you want a Mercedes, it's not because of the engine they have. Same thing is true with microcontrollers. We have -- we don't only have ARM and MIPS. We have ARM and MIPS and PICs and AVR, and we have lots of different architectures we use. And we largely are successful because of brand of a microcontroller. A customer buys microcontroller from Microchip because what the Microchip name means, the service they expect, the long-term delivery they expect, no obsoleting the part, not forcing them to redesign. Only passing cost increases to them and not doubling the price, taking advantage of the environment. So it's really the relationship with the customer regardless of the core. I think we're doing well everywhere.

On the total system solution strategy, the idea here is that over time with the portfolio, right, systems focus, you're going to drive more dollar content or attach per customer opportunity, right? And I know -- and I keep asking you this every conference that you attend. And I know it's hard to quantify, especially after the Microsemi acquisition. But can you at least tell us if dollar content opportunity or dollar content per program or opportunity is at least trending up year-over-year due to TSS?

Absolutely. The dollar opportunity is trending. The number of parts per design is trending up. So all those indicators are really looking well and up and to the right. As I mentioned earlier -- I think I mentioned a few times before in various places, is that we have the product portfolio to take the entire board today, take the microcontroller, take the power management, take the voltage regulators, take the study plan, take the flash memory, take the USB, the Ethernet, the WiFi, the Bluetooth, whatever else the customer needs, the entire enchilada, and to get that design win. The challenge now is you win that design, and then the design goes to China, to a subcontractor and subcontractor takes some of those parts and redesigns them out to their favorite supplier or where they can get the parts cheaper or somebody who's willing to make a lower margin on a Chinese supplier or whatever. I think that is the challenge that we're getting better and better at tackling because when they provide support to the customer, in our entire solution, there is -- always there is some software in it. So we have learned to have a click-through license on that software. And if a customer doesn't use one of our part, then they're no longer entitled a license on that software. If the customer breaks the bundle and take some parts from somebody else, then they have to pay the price, which is some of the parts, which is higher than the bundle, the whole. So there's a penalty. And therefore, the advantage of displacing us on some commodity part is not there because you're going to pay the same price. So we're getting better at that, and it doesn't work every time. There is some friction in it. And I think that's where the learning is happening. And as we're getting better and better and better, I think we're going to get better and better outcome out of larger and larger designing around our microcontroller.

Yes. I went to one of your major distributor websites recently. They listed over 3,000 reference designs for Microchip. Some of them are -- yes, some of them are subsystems, right, as a part of a bigger system. But some of them are for full-blown systems, targeting a whole bunch of different end market applications. How effective are reference designs in helping to drive TSS? And does the Microchip team have an active program where you're developing new reference platforms for potential customers?

I think that's the #1 most effective way where we have reference design pretty much for everything. We can show a customer how our part can make a digital thermometer to a module for a refrigerator, to a console of the car, to a digital thermostat to -- you mentioned 3,000 reference designs. We clearly have a reference design for everything. And the new ones are coming all the time, using newer products, adding connectivity, connecting to Internet, using newer protocols, whether it's Bluetooth or LoRa or whatever. So reference design is really numero uno. But beyond reference design, there's also -- when we're engaged with a customer for which we may not have a reference design, when we ask the customer, "What are you building?" And customers says, "Well, I'm trying to design XYZ." Then we have talked to sales force to ask the question, "Is it going to be connected to Internet?" Customer said yes. "How will you connect it? Ethernet, WiFi, Bluetooth? Mark it down. Is it plugged in? Or are you using battery? Yes or no?" So as we get those questions, we know what to sample. We know that the customer will need a WiFi chip or a module. We know the customer will need a power converter. We know they will need a voltage reference, we'll know they'll need power management, we'll know they'll need a Flash memory. So by asking those questions, we've gotten really good at it. The people have gotten good at it. So a place where we may not have the reference design, we're able to sample the customers, the entire need of the customer. At the time we're talking to them about the microcontroller, and usually because microcontroller or the central chip is the first one they're trying to decide on. And then they come back 4, 6 months later and say, "Well, I need all these drag chips." But if we've already gave them all those drag chips in a sample, the probability is much higher that they won't look outside.

Let's just turn to the analog business. Team drove strong quarterly growth in analog, looking to do this thing this quarter. You've got a very strong portfolio, right, high performance, mixed signal, power, power management products. But almost pleasantly, it's a surprise to see that the Microchip team has a pretty broad family of silicon carbide-based power modules and power transistor products, which obviously, as we know, are very important for things like electric vehicles and energy conversion applications. Did the Microchip team develop the silicon carbide technology and devices in-house? Or was this acquired or combination of both? And how do you see the growth potential here?

So the start of the technology on silicon carbide came from the Microsemi acquisition, which was now 3 years ago. But a large majority of the products that you see in our portfolio have been developed in our clock, okay? So we have lots of products, lots of solutions, lots of reference designs and lots of design wins in our funnel. But our revenue in silicon carbide is quite small because it's a newer initiative and all those products didn't get developed the first month after acquisition. So since they have been developed in the last 3 months, the funnel size is an order of magnitude higher than really the small amount of business we have. So we're very optimistic about the growth. And as these design wins in automotive and industrial and other places go to production, I think it will be significant growth on the silicon carbide. And we are at -- within the silicon carbide offering, we have discrete power products. We have modules. We have integrated power solutions. So we also have some digital programmable gate drivers. In the entire high-power, power management area, there is 30-plus years of expertise that came from Microsemi, and then we have harnessed it to develop a lot of the products that we're going to -- where we need them. Microsemi didn't have an automotive business, nearly none. And we have a sizable automotive business and industrial business. So we have redirected some of that effort and beefed up those resources and have a great potential, but we don't have larger revenue to speak of today.

Yes. Good to see the expansion of the portfolio on that front. I've got a question from a -- we're going to turn to operations and financials now, but question. So you mentioned your internal capacity additions that have already occurred and are set to occur for the remainder of the calendar year. Can you give us a sense on how much additional capacity you're bringing on in the context of how much capacity you already had? And/or how much extra demand that you are seeing?

Eric, you want to take that?

Sure. So our CapEx forecast for the current year is between $225 million and $275 million. We expect about $80 million of that to happen. That's the midpoint of our guidance in the current quarter. And we've been gradually adding capacity over the last few quarters, too. So capacity is coming on gradually. We have a large portion of our manufacturing that is done outside of Microchip, right? It's done at the subcontractor. So 40% -- about 39% of our wafer fab is internally. About 55% of our assembly and test is done internally. So we can't dollarize for you in terms of the capacity increases, what it means from a revenue perspective. I'm sure that's what everybody would like to back into. But we absolutely are positioning ourselves to increase capacity every month that goes by, and so we're going to be able to produce and ship more in the September quarter and the December quarter than we can in the June quarter and are making good progress on that.

On the -- I appreciate that, Eric. And then on the financials front, I get asked this question, I'm sure you guys get asked this question all the time. On the solid margin expansion profile that the team has delivered, your 70 basis points and 200 basis points gross margin, operating margin shy of your long-term margin targets. Help us bridge the gap in terms of how you plan to get there. Is it more a function of just revenue growth? Obviously, I know that there's some operational efficiencies in terms of back-end assembly and tests, but help us bridge that gap, gross margin to target, operating margin target.

Sure. I'll take that. So our model is 65% gross margin target and 42% operating margins. So we're -- midpoint of guidance this quarter, we're 70 basis points away from gross margin target. And we're about 30 basis points away from the OpEx target. So in total, only about 100 basis points away where the midpoint of our guidance this quarter is 41% operating margin. So we're getting really close. And we're being extremely efficient with our capacity today. All of our underutilization charges have gone away. And we still have some self-help stories in gross margin in terms of internalizing more assembly and test. That is margin accretive as we go forward. We've got some smaller factories that have come to us through acquisitions that we'll be migrating to our more cost-effective factories over time. Pricing is stable and price increases that we're getting, we're passing those on to customers in terms of price increases ourselves. So it's just blocking and tackling from this point forward. We don't have the margins tied to a specific revenue goal. But we're working very hard to show improvement, and that improvement should be gradual as the top line grows. And on the OpEx side, we're making sure that we're investing appropriately in the business and technical support activities for our customers, R&D, new products, new process technologies and support functions to drive the long-term health of the business. So we're getting close to the model, and there will be incremental improvement from here.

And then my last question because we're just about out of time here. But this is on the cash return strategy. I think the team has made it clear on its prioritization of debt delevering towards a ratio of about 3x. But assuming that the team maintains the current pace of debt delevering on a quarterly basis and CapEx funding, I think we estimate that the team has more than enough cash on the balance sheet and cash generation to sustain quarter-on-quarter dividend increases of about 5% -- 4%, 5% per quarter, right, during this period of time as you're delevering. So given the team's focus on shareholder returns, should we assume that the team can maybe continue to just glide path on dividend increases until we reach more of a sort of peer average of kind of 1.5%, 2% type of dividend yield?

I think one thing I would say is that the amount of dividend increase or the percentage of dividend increases is a decision by the Board of Directors every quarter. It's like they haven't made a decision and say, this is what we do for the next 2 years or 3 years. Every quarter, in a meeting, there is deliberation. Those are the kind of increases that I will be recommending. So I'll go in with that kind of recommendation. But I can't promise they'll come out with that kind of recommendation. But I think that's what the intent is, and that's how they're thinking. But I don't want to commit the Board to a given number every quarter.

Okay. Great. Well, we're just about out of time. Always appreciate, Steve and Eric, your insights and giving us an update every time you attend our conference. Thanks for joining us today. 2021 is shaping up to be a pretty strong year for the team, so we look forward to following the progression there. And thank you very much for your participation.

Thank you.

Thank you, Harlan.

Yes. Thank you.