Okay, good morning. I'm John Vinh. I cover semis here at KeyBanc Capital Markets. We're pleased to have Microchip with us this morning. We have Eric Bjornholt, CFO. Welcome, Eric.
Yeah, thanks for hosting me. Appreciate it.
Obviously, just very topical right now is just kind of the cyclical recovery. I think you and your peers are seeing similar recovery trends. Maybe just take a minute, Eric, to walk us through what you're seeing from a recovery perspective, maybe talk about bookings trends and book-to-bill and lead times and customer inventories.
Sure. I'm going to start with the disclaimer that during the course of this discussion, I'll be making certain forward-looking statements. I refer you to our filings with the SEC that identify important risk factors about the company. With that out of the way, bookings have definitely improved. We broke out a book- to- bill ratio in the March quarter. It was 1.07. It was above one in June. We made the comment on our earnings call last week that our July bookings were the largest bookings that we've seen in three years. Bookings have improved, obviously coming off a low base after the inventory correction. March marked the bottom for us. We have seen an improvement in bookings. We had a really good quarter this last quarter, growing 10.8% sequentially and seeing improving metrics in gross margin, operating margin, etc.
Great. Maybe just a couple of follow-ups there. Given that July bookings were your largest in three years, your guidance into the third quarter was kind of largely seasonal. If the bookings were the largest in three years, you know, why not a more optimistic outlook there?
I would say that our guidance for the quarter was above seasonal. You know, we're guiding 5.1% up. We don't have any quarter that is seasonally up 5%+ . You know, bookings obviously age out in time. We have a high level of inventory today. We have extremely short lead times for most of the product portfolio, so there's a lot of turns involved to get to the number that we guided to. You know, the bookings dry up the further you go out in time, right? Visibility is pretty low, and that's just a function of where lead times and inventory are today.
Okay. To that point, I'm wondering if you could just clarify this point that, you know, I think Steve had talked about on the call how, you know, you're sending out letters to your customers and encouraging them to book further out. Typically at this point in the cycle, you're going to potentially see an extension of lead times. There's maybe some concerns that there's maybe some elements of that PSP . Can you just clarify exactly what you're trying to encourage and what you're trying to do with that sort of notification to customers?
Okay. Actually, the only notification that we've given to customers was through our earnings call. In that call, Steve, our CEO, said, "Hey, we've got pockets of lead time extension at this point in time." I think you're referring to historically, we used to post on our website, "Hey, customers, get your orders in because lead times are extending." We haven't done that as of yet. We are seeing, you know, we outsource roughly 30% of our assembly and test. We're seeing that some of that capacity is pretty tight right now as some of the OSAT suppliers are working on iPhone builds and maybe data center AI activity that is pretty hot right now. That capacity is getting consumed and lead times have extended.
It's actually at the point where some of our data center business could be stronger this quarter, but we're not able to deliver what the customers want because those lead times have stretched. That's typically how these things work, that it starts to get lead time extension on a portion of the product portfolio, lead times start to extend, and then you start over time getting better backlog visibility. What we're trying to prevent is customers getting surprised. You have 10 customers come in over the same product, and the lead times go from six weeks to 12 weeks, and they can't get the product when they need it. It's just general communication that we're doing.
Okay. Have you gotten any sort of feedback from your customers since you've made that sort of announcement on the earnings call?
It's early, quite honestly. I haven't gotten any specifically. Obviously, traveling here this week, I haven't got feedback from the sales team. I think customers will take notice and we'll highlight these things to customers, particularly those that are buying those products that are seeing the lead time extension and hopefully get some traction with better backlog coverage. There's really no penalty for the customer to do that. They can cancel or reschedule that product anytime before 45 days before the delivery date. If they gave us 16 weeks of coverage, they still have flexibility to push things out or pull things in.
Okay. Is 45 days the same cancel window that you had previous to COVID?
It has changed over time. I think at one point in time for some of our products, it was as short as 30 days. It extended out to be like 90 days during COVID, and now it's back to 45. I think that's a good spot for us to be.
Okay. Obviously, your June quarter results were extremely strong. Can you just talk about how much of that strength that you saw in a quarter was due to tariff-related pull-ins versus just, you know, recovery?
We think it's mostly recovery. We go out every quarter, our sales team, and talk to customers and have our distributors talk to their customers that had any significant change in their quarter-on-quarter or year-over-year revenue and get that feedback and say, you know, what's happening, right? Most of that feedback is that, hey, our inventory is drying up and we're needing to start purchasing more in line with what consumption is. We did get a few customers, particularly in China, that did note tariff-related. When we expand that out, we think it was probably a $5 million- $9 million impact on the quarter of pull-in.
I think there's an equal amount of customers that are kind of frozen in their tracks waiting for this whole tariff situation to get resolved and understood because, you know, once they buy our part, then they have to manufacture their good, whatever it could be, you know, washing machine, automobile, etc. They want to know what that final landed cost is going to be, and that's difficult when things are in flux. I think it's on both sides. We think it was a pretty immaterial impact on the quarter.
Right. It sounds like you feel very confident that the inventory levels that your end customers are largely normalized. Obviously, if you look at distribution inventory, it's easier to get your handle on that. Can you just walk us through what the process you use because you have so many customers out there to get a sense of where the end customer inventories are and why you have confidence that things are normalized there?
It is challenging, right? We do 47% of our business through distribution. We get real-time reports from our distributors every month that tell us what their inventory levels are. We obviously know what they're shipping through to customers and we know what we're selling to them. We can measure that pretty easily. Distribution inventory has been coming down dramatically over the last seven quarters. The difference between sell-through and sell-in, sell-in being our GAAP revenue recognition, was $103 million in the March quarter. That dropped to $49 million in the June quarter. It is coming down. I think there's still a bit more of distribution inventory to come down, but they're definitely seeing that they've got plenty of products where inventory is kind of at rock bottom and they're needing to order again.
Now, the direct customers, you know, many of these customers signed up for the PSP p rogram early and got preferential treatment. I think there still is some level of inventory that our direct customers are working through. We can't quantify that for you because we don't know exactly what it is. We see a lot of good signs in the business. One of the things we saw this last quarter is that distribution sell-through increased for the first time in eight quarters. That tells me that, hey, the distribution customers are working through inventory and eventually are going to start purchasing more in line with what consumption is. When that happens, the distributors will need to buy to support that activity. The same thing will happen at the direct customers, but we can't quantify it exactly.
There are pockets where customers are running out and there are pockets where they still have some inventory.
Right. When you look at this recovery and you look at the different segments and or end markets, where are the areas where you're seeing the most amount of strength, and maybe what are some of the areas that are still lagging a bit?
We are obviously seeing really good traction in aerospace and defense today. That was 18% of our business in fiscal 2025, which completed in March. Defense budgets are going up around the world. We are the largest supplier of semiconductors to the U.S. Department of Defense, which works in our favor. That is an area that's quite strong. We're seeing the data center business starting to come back, which is good to see. In general, industrial is obviously a large piece of our business, and that's a mix where some customers are buying again and other customers are still working through inventory. Automotive has been a little bit sluggish, which I think is kind of familiar with what you're hearing from other companies.
Right. Aerospace defense obviously has been a pretty consistent area of strength for you, obviously. Can you just talk about your portfolio there? What's differentiating yourself within the market? Is that largely the Microsemi portfolio there?
A large piece of it is the Microsemi portfolio. Obviously, that's a company that we acquired in 2018. We've been working with customers and our product divisions to get more of our products qualified that we can sell those into customers. It's FPGAs that have a heavy footprint in aerospace and defense, but we're selling standard microcontroller and analog products today, some of our memory products, some of the timing products, some of our system level products going to aerospace and defense. It's broad based. It's more than just Microsemi, but that's the largest piece.
Great. You had acknowledged that there was maybe 6% of your customer base where you had some work to do to kind of improve your relationships there. Obviously, some of this had to do with the fallout from the PSP program. Can you just talk about what you are doing specifically to improve on those customer relationships and maybe just give us an update in terms of what kind of progress you're making there?
We have made really good progress. I think at this point, we do not have a customer issue any longer, right? It is going to be the small percentage that have hurt feelings or whatever. We have met with them, fallen on our sword, and explained to them how we got to the point we were at. They understand it, right? Everybody had extended lead times and some form of NCNR program, non-cancelable, non-reschedulable program during the upcycle. We clearly stuck with our program a couple of quarters longer than we should have, and we talked through that with customers. We find ways that we can partner with them to provide value. That is why they have obviously selected Microchip to work with historically. Ultimately, the engineers like working with us. We deal with the purchasing managers and help mend fences there and with the C-suite.
I think we are in a good spot today. Steve announced this as part of his nine-point program. When we went through that work between when he came back in November and our early March earnings call, we had made a lot of progress in that time frame. It has continued over the last four months, and I think we are in good shape. I think customer relationships are healthy.
Right. I think you were also very transparent in acknowledging that maybe the PSP p rogram wasn't kind of the best program to implement during this last pandemic cycle. What have you really learned about it? What safeguards have you implemented in this next recovery? What are you planning to do differently going forward?
Yeah, so, you know, the PSP was definitely not all bad. There are many customers that were PSP participants that would sit here and tell you we had a great experience. We got supply when we needed it. We didn't have lines down situations because of Microchip. There was good that came out of it. It extended too long. We didn't have appropriate safeguards across what we were doing where you could have a situation where a customer has been a steady customer for the last five years and buy 10,000 units a month, right? Fluctuating modestly. The PSP p rogram could have consumed capacity to them where that steady long-term customer couldn't get product for 52 weeks because lead time stretched so long.
We're putting certain things in place where we can look at customer behavior and past history, not just direct customers, but through distribution to make sure that we reserve capacity in the future for those customers and don't let it get consumed when lead times stretch again and backlog goes up because it will in the future. We're using AI/Machine Learning to help us with that to identify those things. I think we're going to be in a much better spot in the future. I don't think we'll ever do a program like the PSP again that goes out 12 months for NCNR. NCNR is important in some instances. I won't say that we'll never use it, but the PSP p rogram, a program like it, will not come back in that exact form again.
Right. As part of your nine-point program or turnaround plan, there is some Fab rationalization in there. What are you doing in terms of just optimizing your Fab footprint to take into consideration the concerns around Section 232 tariffs and some of the geopolitical related tariff risk there?
Right. Okay. We do a little less than 40% of our wafer fabrication in our U.S. factories. It used to be three factories. Now it's two factories. We closed the Arizona Fab. The last production came out of there in May. None of that production that was produced in Fab 2 is moving outside of the U.S. It's moving to our other two U.S.-based Fabs, one in Colorado and one in Arizona. The production in the U.S. for us is not changing because of that. We expanded clean room capacity and equipment capacity significantly during the upcycle. We think that we can support peak revenue plus again from those other two factories because they've got the largest footprint. Fab 2, which we closed, was our smallest of our three domestic Fabs. From a 232 tariff perspective, I think we're actually in a really good spot, right?
We have domestic manufacturing today. We continue to invest in that. I mentioned before that we are the largest supplier to the U.S. Department of Defense, which is pretty important. We're continuing to invest. When you look at some of our foundry partners, we use foundry partners that actually have U.S. footprints also. If you look at the dollars that we source through U.S. fabrication through our own Fabs and then our partners that have facilities here, TSMC has a Fab in Washington, GlobalFoundries has a Fab in New York, X-FAB has a Fab in Texas. You combine those together, we do about 50% of our wafer fabrication domestically, which I think should qualify us to be exempt, but we'll see how the final rules play out.
Right. One follow-up I had on that is, are you supporting any sort of fungibility between some of your overseas capacity with your U.S. capacity? Is that something that you're working on at all?
There is some of that, right? We have certain products that can run in our internal factories, and they can run in an external factory because we have process technologies that are qualified both in our internal FABs and an external partner. Essentially, what we've done is we have mapped out for our customers where they can see this is the point of FAB, this is the point of assembly, this is the point of test, and done that in a way where, depending on where all this tariff stuff falls out, they can pick and choose where material is sourced from that gives them the best answer for them, depending on where we're shipping that product. It is not perfect.
There is lots of product that is produced on advanced nodes where, you know, we do not have the ability to do those by ourselves internally, but our foundry partners might have multiple locations where they can do that, and we will maximize that to the extent that we can.
Great. Are there any questions there?
I just got a random question. This NSA mandated the new kind of architecture. How do you implement that?
That is definitely a new thing for us. We're making investments there and supporting what customers need, and obviously want to be compliant with everything that we need to be on a global basis. It's just continued investment that we need to make over time. I don't think it's a material needle mover for us, but something that's important to a portion of the customer base. I do not know the answer to that. We can follow up separately. If you want to follow up, I can get the business unit person to chime in. Yeah.
Any other questions? Great. Eric, maybe you can talk about kind of the China market right now. Maybe just walk us through your China for China program. Are you also just given the geopolitical tensions out there? Are you seeing more pressure on your Chinese customers to kind of move to, you know, local sources? I know there's always been competition in that domestic market, but maybe just talk about that.
Right. We talked about a China for China plan back in March, and I think with all the tariffs that have changed between then and now and still waiting for some of that to get finalized, that is not a completely finalized plan as of yet. We've mapped out our sourcing of each product so our customers can see that, and we can work with them to find the optimal manufacturing flow that works for them depending on the country of location. China competition is definitely an issue for a portion of our product over time. We do about 18% of our revenue in China. We think about half of that is done through multinationals that we ship product into a foreign trade zone over there. It's manufactured, and then it ships back to the U.S. or Europe. That really isn't at risk.
The 9% that is kind of domestic consumption is something we obviously look at. We think about half of that is products that Chinese competitors don't have, and there really isn't competition for us domestically there, and the customers need to use our product line. An FPGA could be an example of that. A data center product could be an example of that, and a couple of examples there. That leaves probably 4%- 5% of the revenue that is subject to domestic competition. I'll call that kind of standard microcontroller and analog product. Clearly, those customers have chosen Microchip because they want to work with us. There is pressure from the local government for them to source domestically if they can.
That's where our China for China program will come into play and finding a way to make as much of that product be sourced from an area, could be sourced from Taiwan. Is that considered China or not, right? China would probably think so. Taiwan would not. There are a lot of permutations and combinations, and we will work with our customers to best service their needs. There is going to be more and more competition on that 4%- 5% of revenue, and we'll continue to invest and support our customers and invest elsewhere where we think we can get the biggest bang for our buck.
Right. Can you talk about the pricing environment right now, and maybe also just comment on how does that compare to the pricing environment you're seeing in the China market?
I would say pricing for Microchip is pretty stable. We are being aggressive at the point of design, which is normal for this point in the cycle. All of our competitors have excess capacity and are fighting for sockets just like us. We're being aggressive with pricing there. We talked about likely a mid-single-digit pricing decline this fiscal year, and it probably gets more stable after that. That's really just being aggressive at the point of design. Once we're designed in, those prices tend to hold through the life of the application. All the work is done up front. The pricing is set there, and then the customer is designed in with a proprietary product that is not pin for pin compatible with any of our competitors. Pricing is stable.
Right. As part of the turnaround plan, I think you talked about adding an AI business unit. What's your AI exposure today, and what are the opportunities that you see to grow that franchise?
Yeah, this is new, the AI business unit. We created an edge AI business unit that is really looking today at maximizing the benefits of the portfolio that we have today and building a go-to-market strategy around that. Getting all the business units that support the AI ecosystem to work together for common sets of IP, common go-to-market strategy, feeding our sales force with the proper collateral material to make ourselves successful. There's definitely IP investment that's going to be made. There's helping to shape new product development for some of these business units that have the opportunity to really do well there. It could be MCU, MPU, some of our wireless products, our data center products, etc. With that, it's a relatively small portion of our business today, but it's got a growing footprint.
We are very well positioned with some of the products that we have today, just in kind of enhancing the software, working with the customer to do that, to do really well.
Right. Can you talk about the puts and takes of how we think about the gross margin expansion as we continue through this recovery?
Sure. Our long-term target model on gross margin is 65% on a non-GAAP basis. The last quarter was just over 54%. We're guiding the current quarter to 56%. There are two significant charges that are impacting gross margin today. We have underutilization charges from our factory, which were just over $50 million last quarter. We also have accounting charges for inventory reserves, which fell last quarter to about $77 million. It was $90 million a quarter before. Those are going to come down very rapidly as we move ahead. The basis for those calculations is you take a snapshot of inventory and look at 12 months of trailing revenue demand. We've been in a falling revenue environment, and inventory has been, up until the last couple of quarters, pretty stably quite high. Inventory is coming down rapidly.
We've got a goal this year to reduce inventory on the balance sheet by $350 million. That alone will take those charges down. With the revenue inflecting upwards, that's also going to take those charges down. Those charges never go to zero. There's always some inventory that is subject to those charges. That $77 million could come down to $20 million. That's $57 million of gross margin improvement or five percentage points that will come back to gross margin pretty rapidly here over the coming quarters. The capacity underutilization charges will be more gradual. I think that will spread over a couple of years. That charge is going to come down just modestly in the current quarter. That is really from our backend factories, ramping assembly and test activities. We drained finished goods last quarter. We don't want to do that again this quarter.
We indicated publicly that we plan to increase wafer starts in our factories starting in December. We'll gradually do that over time as revenue and inventory supports it.
Great. Looks like we're out of time. Thank you, Eric.
All right. Thanks, everybody. Thanks, John.