Microchip Technology Incorporated (MCHP) Earnings Call Transcript & Summary

Good afternoon, everybody. My name is Raji Gill. I'm the semiconductor and auto tech analyst here at Needham & Company. We're very pleased to have Microchip with us today at our Fifth Annual Automotive Tech Conference. Presenting for Microchip is Eric Bjornholt, the CFO; as well as Matthias Kaestner, Vice President of the Automotive division for Microchip. The format of the session will be about a 15, 20-minute presentation by Microchip, and then we'll kind of move into Q&A. For folks that are on the line, if there's any questions that you might have, there's a chat box, I'll be monitoring it, just feel free to answer the questions there. So with that, let me hand it over to Eric and Matthias. Welcome, guys.

All right. Thanks, Raji, and hello, everybody. Before we begin today, I wish to remind you that in today's discussion, we will be making some projections and other forward-looking statements regarding the future financial performance of Microchip. These statements involve predictions, and actual results may differ materially. I refer you to the company's filings with the SEC regarding some important risk factors about the company. So if we can move on to the agenda slide. So I'm going to give a brief business update today, state of the business and then, really turn it over to Matt to talk about Microchip and automotive, we'll talk about the automotive market. And trends in connectivity, electrification and autonomous as well as the human machine interface revolution. So with that, let's move to the business update. Next slide. Okay. So just a summary, and Microchip had extremely strong bookings activity in our March 2021 quarter, which closed out our fiscal year, and that has continued into the June quarter. And the supply strengths that we have been experiencing have continued to be widespread, and the gap between supply and demand has continued to grow. Our Q1 fiscal '22 net sales guidance that we gave back on May 6 was for the business to be up on a revenue basis between 3.5% and 7.5%, so midpoint to 5.5%. We expect to have record non-GAAP gross margins between 64.1% and 64.5%, and record non-GAAP operating margins between 40.6% and 41.4%. From an EPS perspective, we're guiding non-GAAP diluted EPS for the quarter to be between $1.85 and $1.95, which would be at the midpoint, another record for Microchip. And the quarterly dividend, that was declared back in early May and just paid this past week, was a record $0.153 per share as we've been increasing our cash return strategy to our shareholders with pretty healthy increases the last 2 quarters. Now Microchip is still very much focused on paying down debt. We paid down about $370 million of gross debt in the March quarter. And over the last 11 quarters since we've acquired Microsemi, we've paid down $3.61 billion of debt. So the leverage is coming down. We are really laser-focused on becoming an investment-grade-rated company. Positive movement on that front over the last couple of weeks with Moody's moving us to a positive outlook, which was great to see. And we also just refinanced a $1 billion bond on the balance sheet, which the one that we refinanced -- repaid had an interest rate of about 3.92%, and we tapped into the market for a 3-year senior secured note with an interest rate of just under 1%. So really a very good debt transaction for us that we just executed on. Next slide, please?

Oops. I'm sorry.

That's okay. So we've been getting a lot of questions on our preferred supply program, which we call PSP. And that was a program that was launched during the March 2021 quarter in response to customers' inputs on how we can help them with the supply constraints that they're facing in the current environment. So under the program, customers get priority of supply beginning 6 months after their initial order in the program if they are providing us with at least 12 months of noncancelable and non-reschedulable orders or backlog. And the response to the program has been very positive. About 50% of our backlog today is in this PSP program. It gives us a very solid foundation to prudently acquire raw materials, invest in factory capacity and hire employees to support our factory ramps. And as the supply-demand imbalance subsides over time, which we don't see that happening in calendar 2021 and really expect to continue into -- through the first half of 2022, we believe that some customers will still choose to continue to be in this program, particularly those with good long-term visibility of their demand. And just one note for automotive-specific, given the topic of this conference today, is approximately 80% of our automotive backlog is now in this PSP program. So that's really the update that I wanted to supply today, and I'm going to pass it over to Matt to walk through the automotive content that we have.

Thank you, Eric, and good morning and good day, everybody. I'd like to give you a quick overview about the automotive market and the COVID-19 impact on that market. The data is from a company called MarkLines. They're collecting world production data listed here by month. So during the first COVID wave last year in March, April, May time frame, starting in February in China already, we saw a very significant reduction of automobile production around the world. And it started to recover in the June, July time frame and then going into full swing again in the September, October, November time frame, hitting about 8 million units a month. During that time, semiconductor buffers all along the supply chain, be it with our customers, with distributors and our manufacturing chain got depleted. And as orders were canceled during the April, May time frame and only came back in much later, we're running now into the IC shortage, resulting in lower manufacturing numbers starting beginning of this year, and this is not over. So it's a strong IC market, combined with the cancellation in late orders, the steep recovery and the lean mantra of the automotive industry, leading to the severe imbalance of supply and demand that we're working through. If you look what is driving the automotive business, there are basically 3 automotive megatrends, connected, automated and electrified. And each of those megatrends ties into megatrends that Microchip defined or identified, including 5G, IoT, data center, artificial intelligence, electrification and ADAS. I'd like to cover a couple of them, just give you a few glimpses what this means to Microchip. So in the connected area of a car, 20 years ago, no car was connected to the outside world. Now everything is getting connected, that's requiring also new architectures. And the architecture is becoming more and more software-centric. There's really a value paradigm shift. In the past, the hardware, as such the car, was defined by its handling, power, torque, et cetera, so really for the gearheads. In the future, which started already a couple of years ago, the car is more and more defined by the software, including assisted features, autonomous features, the infotainment domain, connectivity, features are adding over the lifetime of the car through software, through over-the-air updates and upgrades and open a new business model for the carmakers, which is service as a revenue stream. All this goes along, with standardization of the computing hardware and a lot of value being added through software. For example, the autonomous software driving stack. If you look into the architecture, there will be centralized compute units inside the car with a lot of processing power, we can call it data center on wheels, connected to so-called zonal electronic control unit, they also possess significant processing power, and they're all interconnected. Today, those interconnections are done with many, many different automotive buses, which makes it complex. In the future, we see a very clear trend that those interconnections will move towards Ethernet, towards PCI Express to interconnect the centralized compute units at higher speeds and to connect to the consumer world with USB. Microchip is strong in all those areas, investing heavily in automotive Ethernet. We have a very significant PCI Express franchise from the data center world that we can bring into automotive. We're also the market leader for USB connectivity inside the car. I just want to take one example, which is on the Ethernet side, there are different speed grades. There's 10 megabit, a new standard for automotive; 100 megabit; 1,000 megabit; 10 gigabit. While the physical layers, the PHYs are all different; the software layers, regardless of the speed grade are the same, which simplifies a lot the way software is being developed for the car and validated for a modern car, reducing the R&D effort and R&D time to get new cars out on the street. The fewer different software architectures are used in the car, the easier it is to bring more complex functionality into the car. I want to use the example of a new standard called 10BASE-T1S. It's a 10 megabit Ethernet standard. It seems slow, but it will replace traditional automotive networks like CAN and LIN that are used throughout the car today over time. Just to give an example, where a 10BASE-T1S application is connecting, in this case, a sensor. So the red dots in the front of the car are the radars -- are the ultrasonic sensors for the parking sensor, all, over a single cheap twisted pair cable running Ethernet frames. It can also connect the front radar, can combine the ultrasonic with the front radar connecting it to the main processing box. In succession, I'll now flip through a couple of slides showing different applications for exactly one and the same technology. Can connect microphones that are inside the car for passenger communication, for hands-free operations, but also microphones outside the car, adding an additional sense for the future for the autonomous cars outside. 10BASE-T1S is connecting sensors for position sensing of the gas pedals, speed sensors, pressure sensors, temperature sensors for the battery pack, et cetera. It is used to connect lights inside the car. Front light, backlight, indicator lights, ambient lighting, there are several hundred LEDs today in the most advanced cars, they all can be controlled through a single node or a single net of 10BASE-T1S. And at each lamp, there will be a node, and this node consists of a microcontroller and some analog functionality and a 10BASE-T1S transceiver. And we foresee there 10-plus nodes per car for lighting applications only. It doesn't stop there, goes into actuators. It can connect speakers, having active speakers throughout the car, replacing the need for a central amplifier, can be used for emergency call units for e-sound modules, et cetera. It enables also windows, mirrors, door openers, cameras, wipers, seats. So all the actuators can be controlled through 10BASE-T1S. Huge number of potential nodes, and it's enabling complete zone architectures. And the zone is, for example, the door itself. And inside the door, you have different applications like the window lift, mirror control, the speakers, the lock, ultrasonic sensors, ambient lighting, indicator lights. And all those different functions can be controlled over a single cable. So huge potential, making it easy because it's all on the Ethernet protocol, there is no conversion back and forth from LIN to CAN to Ethernet to MOST and other technologies. Moving to electrification. This market is really on the fast track. And I think it has been accelerated through the pandemic. A lot of the traditional carmakers are launching a huge number of new electric models. And of course, there are the obvious opportunities, electric motor, braking systems, DC-DC converters, but also less obvious ones, like, for example, the e-sound starting this summer in Europe, it will become mandatory that an electric vehicle below a certain speed is emitting a sound because it's so silent to warn pedestrians. To go into one example, which is high-power EV charging, level 3, up to 250 kilowatts of power. Taking an AC current from the grid, the high-power charger needs to convert AC current to DC current, and efficient power conversion is key. Every percent that is an efficiency that is gained is a percent less energy that is needed to drive and power electric cars. So for -- to achieve the highest efficiency, we offer dedicated microcontroller solutions that are designed for digital power conversion, and highly efficient and fast switching silicon carbide devices, MOSFETs and diodes, that can increase the efficiency. This conversion for high-speed charging can be done -- is done at the high-speed chargers and fed into the battery. If you charge your car at home, you have the onboard charger, and the onboard charger converts the AC plug that you have from your home in your garage to a DC current, which is required to charge the battery. Here are just 2 block diagrams for the high-speed charger outside and for the onboard charger in the car. And each of the yellow -- of the blue boxes is an opportunity for Microchip in here. Let me move to the third area, which is the automated theme. And this is more progression through different levels of autonomy from Level 1 to Level 3. It's rather an evolution than a revolution as the fully autonomous Level 5 car in volume is still many years away. Driving automation today has many different assist systems that work together. They are listed in green. And I want to pick out one, which is not the most obvious one, that's hands-off detection for the steering wheel. So as long as a car is not yet fully automated, the car needs to know whether the driver is still touching steering wheel. And this can be done through a camera, for example, inside, but the most obvious and cheapest way to do it is to put capacitive sensors inside the steering wheel. So that like a touch screen or a touch pad, so that the car knows whether the driver is touching the wheel or not. If you look at the opportunity in an advanced steering wheel, there's power conversion, DC-DC converters. There is haptic feedback. There are microcontrollers that control the touch pads. And in certain steering wheels, there are up to 13 different touch-sensitive areas that are controlled through 2 32-bit microcontrollers with the relevant touch hardware, 2x power management, of course, LIN connectivity back to the car to feedback the relevant information. All that is safety critical. So all the components need to be certified to functional safety standards. And single touch-enabled steering wheel is a large opportunity for Microchip. Let me jump to the last area, which is the human machine interface. And here, we really see a revolution rather than evolution. If you look at a car, a recent car, not too old from 2013, the number of switches you have, even the dialing pad for the telephone was individual buttons for each number. So a huge amount of switches and this has moved very, very dramatically in the course of 7, 8 years, that's the equivalent car today, almost no switches, screens, all of them are touch enabled, with lots of touch surfaces and touch functionality, even gesture functionality included in there. If you break it down a little bit, what the opportunities are and the car you've seen before is on the road today. Let's start with a steering wheel. You have a small display in the steering wheel. We need a microcontroller, a touch pad control, touch buttons, the ambient light switches, and touch buttons for our door operation in both doors. In the middle, and I'll jump back quickly, you see this wooden surface. That's a touchable surface where we have smart surface control, of course, touch screen controller for the main screen, for the secondary screen. And each of those applications is one microcontroller plus power management plus connectivity plus dedicated capacitive sensing firmware that we program into our microcontrollers or help our customers to develop it. So this paradigm shift or this revolution from buttons to touch is tremendous. The even surfaces, where the touch functionality is behind, for example, the fabric and the roof so that gestures were to open the sunroof even without touching the surface itself. It's all capacitive touch technology. It doesn't stop there. If you look, last but not least, we're all human beings. We like tactile feedback. And for some functions, it's beneficial not to have to look at the screen. So haptics are getting added back to the screen, in this case, chose a rotary encoder knob, which is simply glued onto the screen. It is enabled by Microchip's maXTouch software and doesn't require any additional electronics. So the touch controller I see that controls the touch screen picks up the signal from this rotary encoder knob and with special software, we can read out how fast it's turned, how far it's turned, et cetera, et cetera, opening additional functions, and those will hit the road pretty soon. You see we have a pretty broad portfolio of automotive products and can cover many different applications and segments. Rushing through to some of the highlights and interesting areas, I would like to thank you for your attention, and we'll be open for questions.

All right. Great. We can kind of move to fireside chat now with the remaining time we have. Thank you so much, Matthias, for that detailed overview. Eric, if I could start with you. You had mentioned that demand is continuing to outstrip supply, that the gap has kind of widened. Could you talk a little bit about the pace of the orders that you've seen over the last 3 months? And also, can you talk a little bit about where the demand is coming across in certain end markets? Are there certain end markets that are kind of stronger than normal in terms of the order rate growth?

All right. So we're sitting on record backlog today. We've made some public comments about the gap between supply and demand continuing to grow. There was some commentary made on our earnings call that we left the March quarter with about 40% more shipments that we could have made in March based on customers' demand and orders than what we were able to fulfill. We call that unsupported backlog that's requested in 1 quarter and can't be delivered until a future quarter. And that number has grown. We expect it to be about 50% as we end the June quarter. So backlog is growing. The number of customers that we have participating in the PSP program is growing. It's now about 50% of our overall backlog. From an end market perspective, we've been pretty specific that over the last couple of quarters, the biggest strengths have been in automotive, industrial and consumer appliance. But we're seeing the data center business come back nicely. Aerospace and defense took a little bit of a pause in the March quarter with the new administration coming in the U.S., but that business is picking up also. So there's really not any weak spots that I can point to in the business today. Really, all the challenges that we have right now are operationally focused and trying to get as much output out of our own factories and those of our subcontractors to support customer demand.

So there's been news that TSMC recently is talking about increasing wafer capacity by about 60%, getting it almost about where it was in pre-2000 -- pre-pandemic levels, in 2019, actually higher than that for auto. Are you seeing any constraints being eased on the wafer side or on the test assembly for auto or any other particular market?

So we really have not seen that as of yet, and I can let Matt chime in on automotive. But the capacity conditions are still very challenged. You know that bringing capacity on in semiconductors just takes time. And we're increasing our own capacity, and our subcontractors are 2 every month, but it's a gradual process. And I think a lot of the activity that's happening at professional foundries is on the leading edge nodes, which most of our business is not. Anything you'd want to add to that, Matthias?

Yes, no, that's exactly the point. We see many of the automotive chips are using a little bit more mature process technology nodes, not the 5-nanometer, 7-nanometer ones. And there was very little to no investment over the last couple of years, let's say, since 2017 on the foundry side, and those are also the most constrained foundry corridors that we see.

I see. So you haven't seen any kind of movement on the foundry side about increasing capacity, even though, obviously, it's more mature nodes, legacy nodes, there doesn't seem to be any indications that there's more wafer supply coming?

Not a significant amount.

No significant amount. Okay. In terms of the PSP program, and you've talked about that in detail. So it was 44% of the backlog last quarter. Now it's up to 50%. Your -- as part of the PSP program, the customers have to provide you 12-plus months' worth of sales visibility, is that correct?

That's correct. The customers provide us with 12 months of non-cancelable, non-reschedule orders. Some provide us more backlog than that, but the requirement is at least 12. And that's on a rolling basis. So as every week or month rolls off, they need to place more orders with us to keep that 12 months full. And in that, they get priority of supply. That supply priority start 6 months after their initial order date in the PSP program. And so really, that doesn't kick in the priority of supply until around September 1 based on when we introduced the program. But the customer response has been extremely high to this. There's small, mid-sized, large customers, distribution customers that are participating in the program. And it's not a guarantee on price in any way, it's just supply. And it's the best efforts for us to be able to meet their needs. And customers were looking for a solution, and this -- the response has been very high to the program and customers like it.

And we see that more and more carmakers follow that as well as they do not only give an outlook to their customer -- to their suppliers who are our customers but they also give them guarantees that they will take the parts at the end.

So it's interesting coming out of this -- as we come out of this capacity constraint environment, whether it's 6 months from now or later, what kind of changes that are -- do you think are going to -- do you think the company is going to enact? And what changes do you think the industry is going to kind of enact that would be permanent to somewhat prevent the situation from happening again? This obviously was a unique situation because you had demand basically disappeared because of COVID and then it snapped back pretty aggressively. But it is part and parcel of this kind of overarching issue that the industry has moved to more of a just-in-time inventory since 2008. And so our automotive companies going to start -- is the industry, do you think, going to start pushing automotive companies to hold more inventory, chip inventory? How do you think the inventory management is going to change coming out of this?

Well, I'll let Matt speak to what he views, the view on automotive, and I can add comments in general.

I think from an automotive perspective, the OEMs would like their suppliers to hold more inventory on their side. The question is whether they're willing to pay for it at the end? Right now, they're all pro, they are all in favor of it. Just to do it, we give you the guarantee. But it remains to be seen once demand is softer, lead times come down, whether they still remember the crisis or for how long they will remember this current crisis.

Right. Short memory.

Yes. And so do customers, long-term, move away from just-in-time? We're not sure. There's definitely signs that that could happen. But the supply-demand imbalance is still so large we're going to have to see how this plays out over the coming months.

On the PSP program, did you talk about prices? Are you -- what is the price delta in order to get -- to become part of this program?

So the PSP program is not a pricing program. And with that, the customers place the order based on what the price is today, but it is also not a guarantee of price. And so you know that we are seeing cost increases from our supply base, whether it's raw materials, wage costs, foundry, assembly and test services, whatnot, are going up. And so it's very clear -- made very clear to customers that this is not a guarantee on price. We are being fair with customers and not trying to gouge them on price. But as we see cost increases from the supply chain, we are passing those on to customers. And at this point in time, that has gone very well for us. So customers understand that this is a unique situation. And we've been fair with them over time. We're not trying to damage long-term relationships. But obviously, we're focused on maintaining our margins through this process.

So you're working on kind of continuing to expand internal capacity. You talked about spending about $80 million last quarter on new equipment, and it's up about $250 million for the fiscal year. Can you talk a little bit about the lead times that you're seeing for new equipment are very long? We obviously talked about the foundries being, on the lagging edge nodes, that they're maybe reluctant to invest in lagging edge technology. So we have to invest in capacity. The industry has to increase wafers. However, we're dealing with this problem where the lead times to get equipment are long and the lead -- and then there's kind of a disincentive for the foundries to really invest in wafers for mature nodes. How does that reconcile itself from your perspective? It's kind of a catch...

Yes, it is. So our capital forecast for this fiscal year is between $225 million and $275 million. And that's up significantly over what we've seen over the last couple of years when the industry had its challenges with the U.S.-China trade war, followed by the pandemic. But it's still a relatively low percentage of our revenue that we're investing in CapEx. You know that we have a split of both internal and external capacity. We do about 39% of our wafer fab in-house and in the mid-50s percentages of assembly and tests. And we don't really see that the wafer fab percentage is going to change over time. We think that will be relatively stable. And we'll be adding capacity as our requirements for internal production go up just as the top line grows. But assembly and tests, we're absolutely making investments to increase the percentage of that activity that we do in-house, and that's included in our capital forecast for the year. The challenge with capacity is it comes on slowly. We can't just snap our fingers and have more wafer capacity tomorrow. Some of that wafer fab equipment can have a 12-month lead time from order date to when it's delivered. Then it needs to get installed and qualified, and then you start wafers on it, and it's still many, many weeks before you're going to get any production out. So it's a long process, and I think that's a good thing that capacity can only come on so quickly. And one thing with the PSP program that I didn't mention before, is we really think that customers are very thoughtful in the backlog that they're putting on us because it's a firm commitment from them. But as things start -- the supply and demand imbalance starts to come back more into equilibrium, we think we will get early signs through that program if customers start placing less orders on the PSP. So we think that's going to help us navigate this to a soft landing at some time in the future. But the supply and demand imbalance is not going to get fixed anytime soon.

So you view kind of the PSP program as a way to somehow kind of monitoring the risk of potential double ordering that may or may not occur in your business?

It is. If we didn't have the PSP program and we just had kind of a standard 90-day non-cancellation window, we'd have a huge amount of orders sitting in month 4 and month 5 that customers could cancel or push out on us at any time. And with 50% of our backlog about in the PSP program, they don't have that flexibility. And with that, they are, again, very thoughtful in placing those orders with us to make sure that they're putting on what they feel is the right level of orders to support their business 12 months out in time.

So given the shortages, how do you think about the CapEx trending over the course of this year? And then as you kind of think about in calendar '22, it's been 3%, 4%, 5% of revenue. How are you going to manage your CapEx yourself in light of kind of these puts and takes in terms of demand and supply?

Yes. So we still think a relatively good range for us is 3% to 4% of revenue to target kind of long term. And there could be years where it's slightly higher than that. Obviously, the last 2 years have been well below that. But we think that gives us the capability to invest in the business where we need it from a manufacturing perspective. And with that, drive the right level of capacity that we need. And we have about $80 million of CapEx in the current quarter. We've got a $250 million roughly estimate for the current fiscal year. And it's steadily going to come in and help us increase capacity as we proceed through this fiscal year.

Your margin improvement plan is currently in progress. You've seen kind of movement on the margin on the positive side. Kind of which changes have had the most meaningful impact thus far? And what other changes are you targeting for further improvements?

So we've had pretty significant gross margin improvement over the last several quarters. And the biggest driver there was removal of our underutilization charges that were flowing through cost of sales. And now we're running our factories as full as we can. So we don't have those underutilization charges anymore, and guiding at the midpoint this quarter to a 64.3% gross margin and a long-term target of 65%. So we're not that far away from the long-term target. I kind of view the activities that we have now to get to that 65% as just kind of normal blocking and tackling. We're going to continue to invest in the business, expand our operations within our existing factory footprint, which always makes us more efficient. We're going to bring more of the assembly and test operations internally. We have some longer-term projects in terms of shutting down some of the older factories that we've acquired through acquisition that are less efficient and moving those into our more highly efficient factories. But these things all take time. But when investors are modeling out our gross margin from this point forward as the revenue increases, we expect gradual increases in gross margin.

Good. I think we're going to have to leave it there. We're hitting our time threshold. Thank you so much, Eric, and Matthias. As always, I really appreciate it, and thank you, everyone, for joining us today.

Thank you, Raji.

Thanks, everybody.

Goodbye.