Great. Good morning, everybody. My name is John Vinh. I cover SEMIs here at KeyBanc. I'm pleased to have Microchip here and Eric Bjornholt, CFO. Welcome, Eric.
Thanks for having me. Appreciate it.
Sure. Well, maybe where we can start, right, what I think is kind of front and center, everyone's just thinking about the cycle. Obviously, we've got a lot of market volatility today. But obviously, I think you had stood out where you had talked about seeing some green shoots in kind of the first quarter. Things have kind of seemed to be a little bit more muted here in the second quarter. Can you just walk us through kind of what you had initially observed there in terms of these green shoots and how things had kind of played out as we kind of moved into the second quarter?
Okay. Sure. So I'd just start by saying during this discussion, I'll be making certain projections and other forward-looking statements about the future financial performance of Microchip. I refer you to our filings with the SEC that identify important risk factors about the company. So with that out of the way, so what we saw dating back to really starting in February is we saw our bookings were increasing. And we saw that February was higher than January. Then March improved again. April improved again from that. Then we gave guidance for the June quarter, which we met, but also had indicated that we'd seen these green shoots. And it wasn't just bookings activity. It was fewer requests for cancellations and push-outs and requests from customers to actually pull in orders that they had placed out in time to an earlier date.
So we get to May, and May bookings were pretty flat to April. They improved just modestly. And then June was kind of flattish, maybe a little bit up from May. But what we were really expecting and what we see in most cycles is when these green shoots start to happen, that there's an acceleration. And that acceleration just really did not occur. The rate of bookings that were coming in, all they were better. And I should note that the June quarter bookings were almost 50% better than what they were in the March quarter. But the momentum has kind of paused where bookings have been relatively flat and not increasing at that rate. Now, we still are seeing that the cancellation and push-out requests are moderated. We have less backlog.
Our PSP program, which we might talk about later, has been kind of unwound, and no orders have been placed for that for quite some time. So that could explain some of that. But we are still seeing customers that have orders placed out in time, could be out in October, and now saying, "Hey, I need that order in September." So that's not really a true turn order, meaning a booking that turns to revenue in the quarter, but it's a booking that was out in time that now the customers are pulling in. So the green shoots are still there, but they're probably not as green or growing as fast as we would have originally anticipated.
Great. And then, I don't know, can you talk about what you're seeing in July versus June at this point in time?
Yeah. I mean, July bookings were relatively flat to June.
Okay. Geographically, I thought it was also interesting. You'd talked about seeing weakness in North America and Europe. I think many of your peers had commented similarly. What about Asia and China? Are you still seeing a little bit of a recovery in this region?
Yeah. So Europe is probably our weakest end market, but both Europe and America were down pretty significantly in the June quarter. And the Far East was pretty much flat. I think it was down about 0.5%. But what we saw is China was really the first to lead us into this down cycle. And they went in first, and it's very likely that they will come out first. And so they were better performing. I'd say all geographies are weak from where traditional performance has been. But China went into it first of the down cycle and probably will emerge first.
Okay. Maybe just following up to that point, it seems like what's unique about this cycle versus other cycles is it seems to be a little bit asymmetric in terms of what end markets are correcting first and what regions are correcting first. And some regions, obviously, are seeing it a bit later. I mean, I'm just wondering what your thoughts are right now in terms of what the implications are for the recovery, right? Could we see maybe a more asymmetric recovery as well? Maybe it's a longer, more extended and more gradual recovery. But when you kind of think through differences in this cycle, what are your thinking thoughts about how things recover from here on out?
Yeah. So what we've seen the last couple of quarters has been relatively consistent, but most end markets are quite weak. The exceptions to that for us have been our aerospace and defense business, which we report as part of our industrial business. And that's about 11% of our revenue last fiscal year. And that has been growing last year, and it's been relatively steady. And then anything AI data center related has been stronger, and that's not news to this room. But we quantified that the AI portion of our business last quarter, we quantified it for you, that it was a little under 5% of the overall business. So that's not going to drive a recovery in total. Automotive and industrial have been weak, and those are large end markets for us.
Positive sign now out of the non-AI portion of data center for us is that it is actually expected to grow not only in September, but in the December quarter. So that's a positive. But industrial and automotive are still weak for us. And industrial is by far our largest end market at 43% of revenue.
Okay. Are there any areas of industrials outside of aerospace defense where maybe you're seeing things kind of further through the cycle or maybe further through in the bottoming process? Because I think more broadly speaking, you've seen more kind of mixed commentary from your analog peers on the industrial end markets.
I would say industrial outside of aerospace and defense is pretty weak for us. I mean, there's pockets of strength. Sustainability, we report as part of our industrial market, at least portion of sustainability. There's pockets of strength there. Things like factory automation, metering and measuring, process control, those things are all generally weak. It is our most fragmented end market for us. And so it's hard to get great read and be able to break that out in detail for you. We only break out our end markets once a year and don't really track it on a quarterly basis. It's very difficult for us to do.
Okay. Any questions? Great. Hey, Eric, I assume right now you're spending a lot of time talking to your customers and just trying to figure out where they are in terms of just overall inventories. I think my question for you on the inventory front is when you talk to your customers at this point in time, how are your customers thinking about what the new norm are in terms of what the normalized levels of inventory to hold? If you compare that to pre-pandemic levels, is that the level that they want to target? Are they targeting higher levels or kind of lower levels of normalized inventories?
So it varies customer to customer, but I would say in general terms, customers are managing inventory very tightly, right? Interest rates are higher than they were pre-pandemic. Inventory is readily available. Microchip has lots of inventory on the balance sheet. We have extremely short lead times. That's the case for most of our peers also. And so customers are confident today that they can get inventory in short order. Now, if you had asked us that a year ago, what our expectation was, we would have thought that customers learned a pretty hard lesson during the upcycle with this big supply and demand imbalance that you can't treat a semiconductor as a commodity because you can be in a situation where you can't sell an $80,000 automobile because you can't get a $2 microcontroller. And so we thought that that would change behavior. It seems to not be happening.
I'm sure there's pockets of exceptions to that, but customers are managing working capital, and they know inventory is available. And so if I were in their shoes, would I do it differently? I'm not sure, right? I mean, they know that if they place an order on us today, they can get inventory very quickly.
Wow. This seems like we could see maybe a repeat performance of what we saw during the pandemic cycle, maybe to a lesser extent, right? If they're just pulling too little inventory, demand starts to normalize, inventory starts to normalize, right? We could be caught in the same situation again, potentially.
It's possible, right? Cycles in semiconductors happen. So it would not surprise me now. I hope we never see it be as extreme as it was this last go-around, right? I mean, typical lead times for us are 90% of our products having lead times of eight weeks or less. That's normal. And that extended out to 52 weeks+. And that was not a comfortable place for the customers to be in, wasn't a comfortable place for us to be in. So I hope it doesn't get like that, but there will be more cycles. And taking inventory too low kind of makes that happen.
That's great. Maybe we can talk about PSP. Maybe in hindsight, it didn't have the outcome that maybe you had anticipated when you had formulated the program. But having said that, I think you guys acknowledge you're not pulling the PSP program. But I've got to imagine that potentially there were aspects of that program that did help you. I'm just wondering if we go forward and we look at the next kind of down cycle, are there parts of that PSP program that you potentially see yourselves maybe incorporating going forward?
Yeah. So just anybody in the room that doesn't know, this PSP program was implemented when lead times started to stretch. Customers were asking for a program to give them more assurance that they were going to get supply. And so we put in a program that required customers who participated in it. In order to get priority of supply, they would give us 12 months of non-cancelable, non-reschedulable backlog. And it worked great. It worked great in 2021. It worked great in 2022. Customers will tell you they were serviced very well through that program. But the challenge was, how long do you let that program go? And we made adjustments to the program, taking it from a 12-month to a 6-month program back in, I think it was August of last year. And did we make the move too late?
We can look back with hindsight and say yes, but we were still having customers at that point scream at us and say, "Hey, I need more inventory. I need it now." And again, they were doing things to protect themselves to make sure they had inventory and didn't get caught short again. And then we ended the program, any new orders under the program in February of this past year. And are there things about the program that work well? Absolutely. Should have we taken our foot off the gas sooner? I think that's clearly yes at this point in time, but it's easy to say that with hindsight.
Yeah. What were the parts of the program that worked well for you?
Well, I would say when times were very tight from a supply constraint standpoint, customers that participated in that program, they got their orders in early. We were able to meet their delivery schedules. We were able to run our manufacturing very efficiently to produce product and get it to them on time. So that did work well. But then when lead times start to go from 52 weeks to 40 weeks to 30 weeks to 8 weeks, it changes. And that happened very quickly in the environment we went through.
That's great. I think you also, on the last call, talked about kind of the 64-bit MCU market. Maybe just talk a little bit more about this. How big of an opportunity does this represent for you? And kind of are the applications and the opportunities here kind of applications that you previously had addressed and were looking at just kind of the next generation iteration of these, or are we looking at primarily new applications for 64-bit here?
It is some of both. It's some common customers that just need more advanced computing in the applications that they're selling. Microchip is uniquely positioned that now we offer everything from the low end of 8-bit to the high end of 32-bit, now extending into 64-bit. The first introduction of products that we have here are actually microprocessors. We have a microprocessor group within the company. We will be introducing microcontrollers relatively shortly in this set. Again, it's really just to give customers options. Then above the 64-bit, we have our FPGA products. We really feel like we are a complete solution for the customer from an embedded processing perspective. Depending on the compute power that they need, they can choose where they want to go in our product roadmap.
The nice thing about it is we have a common development tool environment that we call MPLAB that the customers can use. So they can seamlessly move up and down our product chain. They can have familiarity with it if they work with it at 32-bit MCU to move to 64-bit and beyond. In terms of market opportunity, we think this probably creates an additional total available market for us of probably $3 billion-$4 billion. But it's very early stages. We just introduced products, and there's a lot more to come.
Great. Maybe can you touch on 8-bit? I think there's some concerns that the 8-bit market is increasingly becoming competitive. What are you seeing on that front?
All the markets that we participate in are competitive. 8-bit, we are really the only supplier who is consistently invested in the 8-bit market over many, many, many years. A new 8-bit product today doesn't really look like an 8-bit product from 25 years ago, right? It's heavy analog functionality on the chip. Again, it's just an option in the portfolio that the customer has. It's competitive. 16-bit and 32-bit and analog are all competitive markets. We aren't really seeing anything new from that perspective. There are certain 32-bit, low-end 32-bit products that can sell into the 8-bit market. We are trying to offer solutions to the customer that provides them an advantage. It might be from a power perspective. It might be from the total functionality that they get from the chip, ease of use, time to market, support that they get.
It's still a very robust area for us. We think there's lots of growth ahead of us still in 8-bit.
Great. Maybe we can talk about China. That's a big topic of interest today with semiconductor investors. How are you guys thinking about kind of the potential China risk? Maybe talk about what percentage of your revenues are in China. And there's obviously a lot of manufacturing capacity being installed in the China market, right? There's obviously a big push by the government to push domestic sourcing. Maybe just kind of walk us through how you think about that and how you're managing that risk.
Sure. So we do 18%-20% of our revenue in China. That's where we ship the product. I think it was 18% last quarter and last fiscal year. So we feel that there's probably half of that that is designed outside of China for use outside of China. So domestic consumption might be in the 10% range of overall revenue. And of that 10%, we think there is about half of it that the solutions that we are providing to the market are unique and just not approachable by China competition today. It doesn't mean that it couldn't be in the future. So it leaves us with maybe 5% of overall revenue. And that's generally standard microcontroller analog product that is subject to these new entrants' competition.
But the challenge for any new competitor when you come to the marketplace and you compete with Microchip, for one of our larger microcontroller analog competitors, is that we've got tens and hundreds of thousands of products in the portfolio, and we give the customer options. And when a new entrant to the market comes with a couple of dozen products to the marketplace, the customer might think, and the competitor might think, that, "Hey, I've got a great solution for this." But the challenge is customers are always evolving and innovating in terms of their development. And if they get stuck and they start designing with a product and they don't have a roadmap where they can go outside of that, they waste a ton of time. They invest a lot of money that's not useful. They have to switch platforms and suppliers.
And it's that consistency of purpose, the support we provide, the vast array of products in the portfolio that really makes us successful. Now, we aren't sticking our heads in the sand and not knowing that China is going to continue to invest. We're continuing to invest too. And we have a ton of customers that love working with our products there. They love the support that they get. And we think they'll continue to use us. But there will be pressure, particularly on state-owned companies, to use local suppliers. And that's fine. It's always been a competitive market, and we're continuing again to introduce new products and put more heavy resources in other areas. India would be an example of that, an area that's growing very, very fast. And we'll deploy our resources to give us the largest return to our shareholders.
Great. Great. Questions?
I just asked about pricing during the pandemic and the lead time pricing. Was there well above kind of the trend line? Do you expect kind of that trend line type pricing or being in an inflationary environment where pricing stays above that trend line for a period of time?
Okay. So the question for those listening in is on pricing and whether the inflationary pressure that caused Microchip and our competitors to raise prices during the upcycle, whether that holds, whether that inflationary pressure continues. So we are not seeing a need to raise prices on our customers today, but we are holding prices flat. And we were quite fair with our customers during the upcycle, and we were passing on these inflationary costs. Wages was one of those. Capital was another one. Everybody was buying new capital to expand their factory capacity. For us, where we do a lot of 8-inch production in-house, there was no used 8-in equipment on the market. So we were investing in new equipment. In some cases, we were having to put 12-in equipment in our factories.
Those are permanent cost changes to our structure that will depreciate over many years. Wages aren't going down. So pricing is flat. Now, pricing is always competitive at the point of design, and it is very competitive today. We lead with new customers with our new, most innovative, cost-effective products and are very comfortable being what I would call aggressive on price in those situations like we always have to win market share. Then the margins on those products actually increase over time as they gain manufacturing volume and solidify themselves in the market. So products tend to have a higher gross margin the older that they get.
You mentioned the opportunity with data centers as opposed to AI. Where are data centers just generally? Back end, front end, production end?
Okay. All right. So the question is on Data Center and where outside of AI or Data Center in general does Microchip play? And yes, we have PCIe switches, Data Center Interconnect. We're heavily involved in retimers. So we have a pretty broad portfolio: security, Root of Trust, timing products, power products, thermal products. So we've got a few slides on our website that can direct you to those details, but it's a robust market for us. It was about 18% of our overall revenue in fiscal 2024, our last completed fiscal year. And the AI portion of that was a little less than a third. So a little less than 5% of our revenue is AI, and the rest of Data Center is kind of the normal types of things you think about.
Okay.
Great. Maybe just following up on the pricing question, your gross margins are under 60% right now. How much of this is a function of just competitive pricing in the market versus underutilization? And then longer term, is 60% still the right target?
Okay. So we hit a high in gross margin of a little above 68% on a Non-GAAP basis. That is still our long-term target, is 68%. Nothing has changed there. Pricing is not really impacting margins today. It is more a factor of utilization of the factories and also the fact that we are building quite a bit of inventory. We've had pretty consistent levels of dollar inventory over the last 5 or 6 quarters at about $1.3 billion, but revenue has gone down significantly. So inventory days is inflated, and visibility out in time is pretty challenged. So in addition to taking some underutilization charges in our factories, which was about $36 million last quarter, impacted margin by about 3%. We also are taking pretty large reserves for inventory reserves. We are building product that we believe will sell eventually, but there's not visibility into that today.
Our accounting policies, which are pretty conservative, require us to take reserves. In the future, when we get orders on that product, it will provide a tailwind to gross margins. So that would probably be the first thing that we would see would be reduced, would be these inventory reserve charges that are impacting gross margin. Then as we start to increase factory utilization, as the revenue environment gets bigger, those things roll off a little bit more slower. Clearly, we are building higher cost inventory on a per unit basis today because we're building less in the factories, and that inventory will sell through at a higher cost. So margins have maintained pretty well.
In a very, very difficult revenue environment to be producing around 60% gross margins and around 30% operating margins, it's not where we want to be, but it shows the resiliency of the profitability model and the cash flow that the business can drive.
Okay. Outside of the reserves on this higher cost inventory, as we kind of get through this, is 70%-75% still the right incremental all-through for you tomorrow?
Yeah. We don't give out specific percentages because it will depend on the pace that the revenue comes back and the mix of product that is sold, whether it's something that was previously reserved for, whether it's a really high-margin new product, or something that's more standard margin. So it fluctuates. I'm not uncomfortable with using something in that range, but we don't give a specific target to use. But we feel that margin will show really nice improvement, and we think that we can claw our way back to that 68% long-term target.
Okay. Just last question. You've been managing OpEx pretty tightly here. If we continue to bounce along the bottom here, are you going to be able to maintain OpEx at these levels?
Yeah. So OpEx is down pretty significantly. So our peak quarter of June of 2023, non-GAAP OpEx was, I think, about $465 million. This quarter, we're guiding to about $350 million. Maybe it's $351 million, but it's down significantly. And we've done that reduction without laying off people. We have our employees on a pay cut today. Bonuses were high back then. They've gone to zero. We're managing discretionary expenses as tightly as we can. Now, we need to be careful that we don't do any damage to the business long-term by keeping employees on the pay cut. And when we've done this in the past, it's tended to be in place for about six months or so. Ganesh is doing a communication meeting with our employee group today and speaking to them that there is the end of the salary cuts in the future.
Now, the executive team, the higher-level employees can stay on that for longer, but I would suspect that sometime in the December quarter, salaries will go back to 100% for our employee base. They've already been on the cut for about six months. And so we haven't really ever taken it to that. I think we can go to nine months. We'd put it kind of in the middle of November, at least for our employees in the U.S. But after that, we really need to make sure that we don't do long-term damage, maintain our employees. The reason we don't do layoffs is because we want everybody focused on introducing new products, supporting customers, designing product in so we come out of these down cycles stronger.
Our culture is one of shared sacrifice and shared reward, which really it'd be very difficult to do at most companies to have most of your employee population on a pay cut. And people sign up for it. It's a voluntary program in Europe and Asia, and we've got 98%+ of our employees that have signed up for it. So it's a good program. Employees participate in the good times and really high bonuses, and then things contract, they participate in sacrifice. But OpEx will go up probably starting in the December quarter, but we'll do it modestly. And we'll do it in a way where hopefully we can, as revenue starts to grow, show a decreasing OpEx as a percentage of revenue as we get back to our model, which is about 23% on the OpEx side of revenue.
Great. Looks like we're out of time.
All right.
Thanks, Eric.
Thanks, everybody.