Monolithic Power Systems, Inc. (MPWR) Earnings Call Transcript & Summary

Good afternoon, everybody. This is Matt Ramsay from the semis research team at Cowen. This is the 48th Annual Cowen TMT Conference, but it's our first virtual one. So hopefully, it's been a successful first day for everybody, and we really appreciate everybody participating. Last but not least in the day for me in my presentations, I think this is the seventh one for my team, but really, really happy to have Bernie Blegen here from Monolithic Power. You guys know the company as, frankly, one of the rock star growth franchises in semis. So Bernie, I don't know if you have any intro, comments, just to kind of set the tone for the conversation. I've got a bunch of questions, but I'm going to keep it pretty informal, and we'll just have a chat about the business.

No. Take it away.

All right, cool. I just -- it's been, obviously, an interesting time with the inventory correction that the semiconductor industry saw last year. And then just as we felt like we were coming out of that and we've kind of gotten another kick with the COVID-19 situation globally, and it's been quite a topsy-turvy time for a lot of companies, seemingly with the exception of you guys in terms of your financials. So if you could just kind of remind folks about where things were in the first quarter and obviously you have significant operations in China and what went on during the initial parts of COVID-19 where you might have had a little bit -- little inventory coming into this situation and then built back up a bit, and then we can kind of talk about the forward business. But if you just set the stage a bit about how things were in the first quarter and over the last few months.

Sure. Thank you very much. So if you have followed the MPS story, you know that in 2018, we were expected to grow 17% and actually turned in a 24% growth. So going into 2019, this would have been like November of 2018, we still had those high expectations. And so we -- our build plan for inventory matched that. As 2019 progressed, you will recall that it was about in December, January, January of 2019, when it became apparent that it was going to be a weaker than normal year. What happened was then our inventory, even if we slow wafer starts, it takes 2 to 3 quarters before you see it in my balance sheet. And so what happened is, is by the end of Q2, our inventories were in excess of 200 days, which is high even for us. So our range that we'd like to be at is between 160 to 180 days. So we worked hard to reduce wafer starts. And by year-end, we got the inventories down to 155 days. During the call where we gave guidance for Q1, we just wanted to tell people that while -- and expectations at that point were that we were going to grow 17% that we had inventory both on hand and in our supply chain, sufficient to cover that. But we could not do a repeat performance of 2018, where we were able to address surge demand. The areas that we're most interested in looking at were the 5G, the data center and the gaming consoles. And in hindsight, that was a luxury of optimism that nobody seems to share today. But what we did is in Q1, when -- around the time of the Chinese New Year, it became apparent that the problems were going to affect the industry. And the way that it sort of worked through us, I don't think we're unique, is that we had what we refer to as the toilet paper syndrome. And what I mean by that is if you went into the store and the shelves were all empty, the next time you went in there, you bought. We have people who became hoarders. Well, our customers have that same concern about being caught short. And so what happened was, was when we finished Q1, I normally have 90% to 95% of that quarter's revenue in my backlog. I had such a high and unrealistic number this time, I didn't even bother calculating it. So we find ourselves in this odd position where you think we should have a fall-off in demand if unemployment in the U.S. is going to be over 20%, it goes that there is going to be a drop in demand at some point. We haven't seen it. In fact, we've seen quite the opposite where people have been overordering. And we do it at a time where we've already declared publicly that we do not have excess fab capacity to meet all possible demand anyway. This is a surge point. So what we've had to do is resort to partial shipments whereas we only shipped 50% to 70% of what our customers have ordered. And if there is real demand that we're missing out on, then it's incumbent on them to escalate either through the sales country managers through the VP of worldwide sales or in some cases, Michael, my boss, has gotten 5 or 6 calls in the last couple of weeks from CEOs asking for additional inventory. So we had a situation where everybody was doing what they thought were the right things, and at the end of the year, our inventory was low, the channel inventory was low and even the customers' inventory was low. And almost going in reverse logic from what you'd expect given the severity of the pandemic, actually, demand has increased quite a bit. And I almost argue that, that might become a new norm of people holding higher levels of inventory in the future just as a risk management or insurance policy against possible supply chain disruptions.

That's -- I think -- no, thanks for the sort of setting the stage there. And just a couple of things for me, a follow-up. The first is, is that something you're seeing more specifically to your customers in China? Or is it something that you're -- that the same type of tactics, partial shipments and overordering, is that something that you're seeing more on a global basis or much more in China?

Well, there's actually -- there's 3 things that's generating this additional demand for us. The first is in consumer, and a lot of it has to do with Internet of Things and home appliances. And in that case, I think that is some trade-related concerns that are China-specific. The second thing that we're seeing, and you saw it when we went from Q4 to Q1 in the comms, is our 5G revenue jumped up about $6 million. And I think some of that was both an acceleration of real demand but I also think that, that particular networking company, which is Chinese-based, pulled in orders out of concern for trade. And then the third aspect that is benefiting us has been the work-from-home phenomena, which is seemingly triggering additional build-outs of data centers. In Q2, you should see a large spike in our revenue. And that's due to memory purchases, which usually precede a build-out of the data centers and notebook sales. And then it remains to be seen how this is going to impact gaming consoles.

Yes. And obviously, some -- a good bit of moving parts there. I wonder -- I just to -- asking this another way, you guys have talked for years and, to your credit, delivered on this for a very long period of time, 10% to 15% sort of outperformance versus your -- the market that you serve. And -- but if you look at -- I would think because of maybe some strengths in your overall business that might be outperforming a bit and also some of the overordering that you've seen, I think you might outperform the industry by double that in the first half of the year. I don't know, just ballpark numbers. And I wonder how comfortable you are with that level of outperformance. And are there -- how do we -- I don't know what the right way to ask the question is. How do we revert back to like a normal outperformance level for you guys? And I'm just trying to -- I know there's a lot of investors that, really, are big fans of your company, but then we see that level of gap versus the rest of the industry, have a bit of concern about how do we get from here to there back to some normalized level of outperformance. I don't know if you have any thoughts on that.

Yes. Again, we're a little bit limited, as I said, based on our ability to obtain inventory to service all possible demand. So you have to sort of build in that constraint, unfortunately, to the model. But I would agree with you that there is an opportunity to exceed that historic 10 to 15 percentage points of outperformance, again, recognizing that we are a little bit inventory-constrained. Something I -- I've gotten away with this for the last 4 years by saying that 10 to 15 percentage points. I have to confess, Matt, there's nothing empirically available to support it. It's not predictive. It's the benefit of hindsight, but it has proved to be a pretty safe rule of thumb.

Got it. I -- just a follow-up. And I apologize, Bernie, for continuing to look to the side. I'm not watching Netflix on the side or anything. This is my interface for audience questions. So people keep pinging me. But somebody -- a couple of folks have just pinged me to ask a follow-up. Any estimate that you've provided folks with what you feel the magnitude of overordering, what that one particular Chinese customer might be just so we can gauge what that impact is?

Well, so that's a hard one for me to answer. It's an appropriate question. It's a hard one for me to answer. I'll tell you why. Back in Q1 of 2019, we had 3 new products that went into production with this customer. One was in 5G, one was in data center and one was in consumer, it was a wearable product. So a company that had historically only been low single-digits growth now went to mid-single digits and in any particular quarter, depending on ordering patterns, can go to double digits. Now here's what is a little bit hard for me to get my arms around. When we started out in that relationship on 5G, we were doing power management for an FPGA accelerator. And that got designed out essentially in Q3 of last year. And I just threw up my hands and said, "That's it. Party's over. Oh, well, wait for 2021 now." And yet the revenue continued. But the composition of that revenue changed significantly. So whereas before it was the power management for an FPGA, now it was this low end point of load. They're like jelly beans. And it's interesting because the jelly beans are like -- they're very foundational. They can be built into anything and you can say, okay, when I've got on a particular reward for a base station, the power comes in at 20 volts, and that's what's needed for the battery management. But then if you want to go to the processor, it has to go to 24 volts. So that's what we do. It's boost and buck, up or down. And generally speaking, with most of our end customers, we actually have to codevelop the solution. So we know what the application is, what the specs, how many units. When you order, that becomes sort of price times quantity. Beginning at this stage, it's very predictable. In this case, we don't know if the jelly beans are used for the transceiver, the base station, the fiber optic, the home office, the back-end data center. It's a mystery to us. So as a result, the ordering pattern and the ordering levels are hard to predict. So we obviously have a forecast because we have to be able to have a build plan to this, but I don't have the same level of visibility than it would for either different end markets or different customers.

Got it. No, that makes sense. And obviously that particular customer, I think, has order patterns, for a number of different reasons have been a bit volatile for a lot of folks in the industry. One of the things that you touched on there was just in describing the product was a topic that was actually in my list, where you guys are addressing different voltage levels, particularly in the data center space. A prominent company in the data center market is now launching a pretty big new platform that will get you guys real revenue in the 48-volt arena here pretty soon. Maybe you can talk a little bit about the opportunity for 48-volt power supply and where you're positioned maybe differently in that market relative to some other markets versus competition.

Yes. So one of the interesting things about MPS, and I don't think it's entirely unique but it is interesting, is we developed this technology that we refer to as QSMod, Quantum State Modulation. It is a more precise level of power delivery. And we have been able to go into different market adjacencies by just making little incremental changes to the R&D. And so that's why we have power management for FPGAs, ASICs and in this case, for GPUs. Now in 48-volt, everybody has always known that eventually, one day, data centers would need to convert for energy efficiency reasons. But we've always been looking for that sign point -- that signpost of when that's going to -- the transformation is going to occur. And the reason is, is that it turns out 12-volt can still be just good enough for most like e-commerce applications. So there hasn't been that -- even though there's one company, in particular, one hyperscale that has been making that conversion, it still has to step it down because the Intel-based processor is still 12-volt. But in the one that you're referring to is for artificial intelligence. Now I'm going to sound like I did just a few minutes ago on the incomplete 5G explanation. So if you're seeing a pattern here, sorry about that. And that's as follows, is, again, if I'm looking at hyperscale for e-commerce, I know that there's 13 million units -- servers out there. I know the size of the e-commerce market. I know the size of the server market. And I know the rate of growth. And it's a very mechanical process to estimating what revenue contribution is going to be. In this case, we don't even know what the final configuration is for the server, so we don't know exactly, precisely what the dollar content is. We believe it's over $100, but we don't know for certain. And then we don't know how large a data center has to be. And then the next question goes, is that particular AI application even going to be successful where they need to have multiple data centers? And if so, how many? So it's not like taking a run rate business and layering in the fact that, okay, Intel is moving from Purley to Whitley and actually have this impact, it's how does a fundamentally new market on a new technological platform evolve over time. So in the immediate, I'm very excited because it will demonstrate that we are commercially viable in this very important market. But as far as the revenue contribution that I see in '20 or '21 has been relatively modest.

Got it. So it sounds like the compute business is largely going to be just driven by the -- I mean it's a powerful upgrade cycle, but it's the primary upgrade cycle rather than -- okay, that makes sense. Another -- jumping around to a different end market here is -- the automotive market, for you guys, has been one that I think for a -- 2, 3 years when we were working together and doing these types of meetings, we've talked about that being a 40%, 50% growth business. Obviously, off a small base a few years ago, but now it's an important part of the company, and I think still has some pretty good potential there. 2019 was an interesting year for the automotive business with China sort of skipping a model year and there's a lot of worry about automotive globally from a units perspective, given the economy, et cetera. But are you seeing trends that might be -- I mean I would say of markets that are -- investors are scared and maybe a bit worried about what near-term demand is going to be? Auto would be one of the -- high on the list, but it sounds like your business in China might be acting differently than that, especially relative to 2019, where there were some challenges. If you could talk a little bit about that, it would be great.

Sure. So if you look at our history -- and Matt, you're exactly right. We have $90 million last year. Of the $90 million, half of it was related to USB ports, pretty low-end technology, important but low end. And what you'll see is that our revenue will be sort of flat, and then it'll have a big stair-step usually in Q3. And what that represents is design wins that we secured a couple of years ago that are now going into production for next year's model year. Well, we have a higher exposure to these greenfield opportunities than most of our peer companies. And as a result, we're not as badly affected by what occurs in the SAAR. We're not immune. So for example, if things go as planned, you'll see that in 2019, Q3, I had my revenue step up. Q4, it was about the same. Q1, it came down just a little bit, but it's lost in the rounding. And then Q2, I'm expecting it to fall as a reflection of most of the plants were closed, they were shuttered for the majority or all of Q2 so we weren't making shipments there. But then in Q3, I think we'll resume production. Now could you make an argument that maybe they don't cancel model year but because they're playing catch up on getting production capabilities and tooling that certain are delayed by a quarter? Maybe. So I think that one of the -- there's a couple of exciting things here that I want to leave you with, though, is -- the first is that, as I said, we had 80% exposure to infotainment, and most of that was related to USB ports. Now the design wins that are going into production include the body controls, lighting and ADAS. So that's very exciting differentiators, we get into a broader number of applications. The second thing is, long term, I believe that the value proposition for automotive remains as high and as attractive as we'd believed. One of the things that contributes to that -- now I haven't heard anybody -- so you have to understand, Matt, I do a lot of sudoku puzzles these days when I'm at home. And occasionally, I just have these free thoughts. And I'm like, "Well, it looks like I'm not going to be flying as much as I used to. And if that -- well, I won't be flying for a vacation either, I'll probably be driving." And if you look at a lot of people who are dependent upon public transportation, they're probably less likely to continue with that if they have an option to get a car. So I think that there's many forces that suggest that automotive is a little bit in trouble, there's also a lot of forces that say that it could remain a very high-value opportunity for us.

Got it. But I guess the follow-up to that is, in this particular time, you're fairly confident in the fact that the Chinese market won't be skipping another year in reaction to some of these challenges?

That is correct.

Then you have the order book to sort of back up that -- got it. Okay.

I allowed for the possibility that you could move by a quarter, maybe because tooling issues of the shutdown but that I don't see anybody skipping a year.

Got it. Now that makes sense. There's always a number of different ways we can go, but I think the consumer business, as you mentioned to start the conversation and your introduction, has had some challenges over the last year, 1.5 years or so. And in some sense, it's a challenge. In some sense, that's the piece of the business, at least the low-value side of consumer that perhaps the investment community cares the least about and will let some of the other more exciting parts of the portfolio drive the business, but it's been a piece that you guys have used to balance your gross margin expansion as you built the company over the last number of years. And you and I have joked for a long time about the cookie jar analogy. Maybe you can give us a little bit of an update about how -- as you manage through this turbulent period, pushes and pulls to the business, how you're thinking about managing the gross margin and what that cookie jar looks like right now relative to those efforts.

Yes. So -- just so that people are aware of what the cookie jar analogy means is that there has always been historically some level of business that's tied usually, not always, to consumer. And we can choose to participate in that business or we can hand it off because they're -- these are not single-sourced opportunities. We just choose not to accept the PO. So that's the cookie jar fund. In 2019, if you look at my gross margin, you'll see it pretty much flat wide. It was 55.6% for I think 4 quarters a row, and then we dipped for the last 2 quarters at 55.5%, same thing, same area. And what we were doing there is we recognized back in January of 2019 that we're going to have, like everybody else is going to have to slog our way through the year. And so what we did is we picked a gross margin level at which we knew we could sustain for consistency throughout. Well, that's different than what we have done historically. What we've done historically is we've had gross margin increase 10 to 20 basis points. And to the extent that in order to do that, we had to bring in lower-margin business to offset higher-margin that helped accelerate our rate of revenue growth. So when I look at the cookie jar fund today, it's on fire, I have too many cookies. Some of it is because a lot of it has to do with this -- the toilet paper syndrome and people are ordering ahead of what they need. But interestingly, there have been a couple of trends within consumer that sort of surprised me. One is there's been a conversion in the standard for IoT and for home appliances. And that's driving a lot of near- and midterm demand. And so we actually had in Q1, and we're seeing it again in Q2, pull-in of these orders. So right now, what we're trying to manage around is in Q3, we'll have this bubble of gaming console sales. And even though their profile, their margin profile has improved significantly because we move to 12-inch and we have a lower cost solution than we did 3 years ago, they're still below our corporate average. So the punch line is, is that we're still as active trying to manage the cookie jar fund and pull in as much low-margin opportunity into Q2 as we can, and it is available and then push into Q3 some of the higher-margin activity to act as a counterbalance against the gaming console sales. Now having said all that, from the simplest point of view is that we want to return to our model of just improving quarterly gross margin, on a sequential basis, 10 to 20 basis points. So in one hand, I've showed you how the sausage is made. The fact is all you have to do to appreciate is that it goes 10 to 20 basis points up.

I think I've actually never asked you this question, but I wonder if you reflect a little bit on the fact that you just described there prioritizing certain margin levels in certain quarters. And it just kind of struck me the fact of how confident that you guys be -- must be and the differentiation of the product portfolio must be to give you the luxury of doing that. I would imagine that a lot of -- you've probably managed in other companies where you didn't have the luxury of being able to do that. And I just wonder -- I mean is that a fair assessment of that? I mean it's -- it must be -- I don't imagine too many of your competitors have that luxury.

No. It's a completely unique set of circumstances. And I'd like to say I had something to do with it, I really don't. I just -- I get to absorb the benefits of having it in place. I do sit in 2 or 3 really boring meetings once a week, where we actually go through it at a very granular level, to sort of fine-tune the dials of product mix. But it is completely unique. And part of that, just to make another point, a side note is Michael, as you know, is looking 3 and 5 years out. And he really does not pay attention, believe it or not, on a quarterly, quarterly basis. He just always assumes, "Oh, there's nothing I can do. The head of sales and Bernie have this quarter wrapped up." And so it's a big high vote of confidence for Michael. But it's very unique to have your CEO who's so forward-looking and almost like we're little lab rats that are cleaning up the details of the quarter.

No. I assure you, what you guys are doing is much more sophisticated than that. But I have one of these little interfaces here running these virtual sessions that has indicator lights on it and they're yelling at me that we're bumping up against time. But it's -- thank you, Bernie, for sharing some insights with us today. I mean you guys have a remarkable business that you're running, and we really enjoy partnering with you in that. So if there's questions for investors on the line that I didn't get to or maybe we need to dig into further, please contact myself or my team at any time or we'll get you in touch with the appropriate people at MPS to get the answers for you. So thanks, everybody. I think that's the last session of the day. Thanks so much, Bernie, for your time, and we'll see you and talk to you soon.

Thank you very much, Matt, and thank you, everybody else. Bye-bye.

Take care.