Eric Bjornholt with Microchip Technology. Thank you very much for being here today. I'm Tom O'Malley, U.S. Semi and Semi Cap analyst. Crazy times. I know you guys have been bouncing around between conferences, but I think a question that I like to lead off with, and I've kind of talked about broadly in the conference is there's this huge AI tailwind, right? $3 trillion in spend plus. You guys obviously attack this in a different way than the data center-centric players. Maybe talk about how you guys attack this and, you know, what's going on with the edge.
Sure. So I'm gonna let Rich answer the question, but I'm just gonna say during the course of this discussion, we'll be making certain forward-looking statements about the future financial performance of Microchip, and we refer you to our filings with the SEC that identify important risk factors about the company. So Rich, talk about Microchip and data center.
Yeah. So we're attacking it. Probably you've seen the last release that we did, our three-nanometer PCIe Gen 6 switch, so we're attacking the market from the PCIe switching standpoint. We've been in PCIe switches for Gen 3, Gen 4, a little bit of Gen 5, and now Gen 6 in a very large ramp. We'll be introducing retimers that match with that Gen 6 next quarter, and we also play within the AI data center in timing, secure boot. There's a number of other product lines that also play.
Mm-hmm.
Within that particular market space for us. So it's a large market for us today and growing considerably over the next five years.
It's really interesting 'cause if you look at analog mixed signal players and their reach into the data center, that is very much a digital forward market in certain instances. And so the players that have existed there, the Astera Labs, the Credos, the Marvells, the Broadcom, they talk about software platform and how important that is on a go-forward basis. Can you talk about how your offering competes with them? You've spoken recently about how you guys have something very similar.
Yeah. We have ChipLink. You know, Microchip was a predominant, a major supplier of Generation 3, Generation 4 PCIe switches. We launched ChipLink, which is the software platform that supports our PCIe switches. It is considered the gold standard within that marketplace. Gen 5, we decided to roll our own SerDes. We were about 12 months-18 months late. We are a minor player within Gen 5, but the ChipLink software suite is still there, and we continue to enhance it. ChipLink is ready to go with Gen 6 release. Then because we've also skipped a generation, where our competitors are all sitting on five nanometers, we've jumped to three nanometer.
Mm-hmm.
On this Gen 6 and this allows our customers to move very quickly to Gen 7, as they finish out their Gen 6 deployments.
So you have opportunity within the data center, and then I would assume also as this proliferates to the edge, the microcontroller portfolio. How are you guys seeing your business develop there? Have you really seen an intersection point? Because, I mean, you have your smartphones today, but then there's gonna be devices that you would imagine come in the future. Those need to be connected. Where are we at in that transition?
You know, on the edge, one of the new groups that we formed was the AI/ML group. Believe it or not, majority of the applications that we see coming in for our customers using microcontrollers on the edge do not need an accelerator.
Mm-hmm.
They're having us develop models for vibration, for wear control, for sound, battery management. All of those we were building a model zoo. They can download a model. We, we import their data in. We train a model, and then we port that onto the particular microcontroller or FPGA that they're looking to use. Where we are using accelerators are on our FPGA products.
Yep.
So if you're looking at a vision system, optical inspection system, that's typically where object detection, where accelerators are used. But for, you know, very basic sensors, temperature, light, no accelerators are really needed on those products.
We've talked to a lot of companies here who have not only the microcontroller portfolio but also radios that are associated with those. We went through this transition, the connected MCU. Is radio a differentiator, or is really the microcontroller the differentiator? And how important is it to have the whole suite of both radios and also the compute side?
I think it's actually both. I think what we're seeing is hardware connections.
Mm-hmm.
T1 Ethernet, T1S Ethernet, that is actually the backbone. So when you look at robotics, in terms of where it's going, Industry 4.0, we're seeing much more uptake on wired communication.
Mm-hmm.
We've taken our Gen 4 PCIe switch and created the 12-lane, 16-lane, 24-lane device. That's being designed into, you know, various companies' reference designs for their robotics platform, cars. You need much higher wired communication in these applications. T1S, T1, two-wire Ethernet is really we're seeing that as a huge growth area. In fact, every customer that we talk to, the beginning conversation is not wireless. It's always about Ethernet and T1 Ethernet and T1S.
One more question on this side, just because we've heard increasingly more on this. You've heard the battle of scale-ups, right? Scale-up architectures, Ethernet, UAL, et cetera.
Mm-hmm.
I would imagine with a switching portfolio with ESUN, you would potentially be able to offer some switching products in the back end. Is that something that you guys have talked about? Is that an area of opportunity over the long run in the data center?
So we offer our PCIe switches are both scale-up and scale-out.
Mm-hmm.
So we're doing both. We also have some Ethernet products where we offer scale-out and have been doing that for quite some time. So we already offer products in both areas and you know, probably the PCIe Gen 6 switch offers, you know, probably the greatest growth right now.
All right. So we're zooming in a little on the portfolio side, maybe back up to more of the macro side. So it's been, what, three years now? We've all been waiting for the restart in some of the cyclical end markets, auto and industrial. I was listening to you guys last week. On the margin, it sounded incrementally better, for sure. You guys obviously narrowed guidance as well more positively. Could you talk to the health and wellness of auto and industrial? Is this the time to say, "Okay, we're, we're off to the races here," or is this still, "Things are a little bit better, but we need to wait and see"?
Yeah. There was so much inventory that was built up during the last upcycle that, you know, we've been with 100,000 customers, we've been, you know, slowly working through that. And it's taken a long period of time.
Mm-hmm.
You know, what we are seeing now is broad-based booking strength across geographies, across end markets. And that is telling us that more and more customers, as we move ahead every month, are needing to buy inventory more in line with what consumption is. And so I think that applies to automotive and industrial. Rich, add anything to it?
No. I think a lot of industrial customers who are down probably about 20% from their peak, 20%-30%. And most that I have spoken to are looking for, you know, just consistent, mid-single-digit growth going forward, you know, sort of re-baselined in our growing from here. And automotive market, you know, most a lot of the inventory is pretty much depleted for the most part there. And we're probably gonna be looking at consumption. In fact, we're seeing, you know, a lot of our automotive customers start to lay out backlog on us now going out six months or nine months from now.
Mm-hmm.
And so it's getting back to more normality in terms of how orders layer in for us.
So really it's a conversation with your customers around their inventory levels, which seemingly are better than they were before. And then bookings are a reinforcement of that idea with customers willing to put orders on you for the next, you know, three, six, nine months.
Yeah. And, you know, our lead times are still very short. You know, we've got 199 days of inventory.
Yep.
Lead times are short. There isn't a huge incentive for customers to give us a bunch of visibility. You know, we're having customers that haven't bought from us in some time coming back.
Yeah.
placing orders and that telling us that we're getting nearer to the end of this inventory correction.
Hey, there's also another factor here. I think some people overlook sometimes, you know, during that whole COVID period, there was two years of essentially no new product design, right?
Yeah.
By most of our customers, right? They were looking for alternative sources. They were trying to figure out how to get products out. You know, and some of those designs that were post that COVID period are starting to come to fruition. We're about six months- 24 months, depending on how you time that, where those new designs start to come in. So we're starting to see new designs come to production and layer in on top of also inventory correcting. So there's another factor in there too.
So I did a little napkin math, and please correct me if I'm wrong, but if I look at pre-COVID seasonality in the March quarter, it's up modestly, 1%-2%. Tell me if that's off. And then if you think bookings are better, lead times are short, we're halfway through the December quarter, I think people walked away feeling like, "Hey, maybe into next year we could be a little bit better than seasonal. I don't want you to guide." But can you talk about, like, marrying those two? Is that the right way to kind of think about things broadly?
Yeah. So I would say typical March is in the range you say. You know, it's flat to up a couple percent.
Mm-hmm.
You know, we think that we can do meaningfully better than that with the way the backlog is layering in for the March quarter. So, yeah, obviously December quarter is ending up better than what we had expected just a month ago. We are seeing backlog build really nicely. So I think we're in a really good place, not just for March but to show a really strong 2026.
Different companies that I cover work in different ways with the channel. Could you describe your relationship with the channel? And then is there any behavioral changes that you're seeing, either with your direct customers or the channel that vary from one another? Would that indicate any sort of various trends? Like, there's been a concern out there that channel would be pulling a little more because customers are taking inventory, harder to read those guys than direct customers. Could you walk me through any of that?
So, you know, we operate with three different types of distributors. We use the global distributors, the Arrows and the.
Mm-hmm.
And the Future Electronics of the world. We use the catalog distributors, like the DigiKey and the Mouser that help seed the market with new designs. And then we have a bunch of what we would call regional distributors that are more kind of direct extensions of our sales force that, you know, might just carry the Microchip product line from a microcontroller perspective at least, and then can be pure advocates for us with the customer. All distributors have been taking inventory down. Sell-through activity from distribution to their customers has exceeded sell-in for quite some time. It was about a $53 million difference last quarter. But distribution inventory is getting to the point where they're going to start needing to purchase in line with end consumption.
You know, we think that inventory that has bled out of distribution, you know, is sitting with the customers of theirs, which we don't have complete visibility into, but we think they've been draining inventory also. That has a nice kind of a double effect that will impact revenue. In terms of our relationship with the channel, it's quite good. You know, about 47% of our business goes through distribution. It's an important way for us to reach the long tail of customers. You know, many customers choose distribution because they provide services, whether it be logistics, kitting, payment terms, whatever it might be. We don't steer customers one way or the other. Distribution is an important piece of our network. I think inventory might come down just a little bit more, but it's getting close to being corrected in distribution.
Yeah. So it sounds like the bookings alone are giving you confidence into the March quarter, but then there might be this second wave of actually restocking some of the channel, which has gotten extremely, extremely lean, is undershipping in demand given where inventory was historically. That could be another layer on at some point maybe in 2026.
It is. And, you know, the bottom line is customers and distributors react to what our lead times are at, right? And if they can call us up today and say, "I need product in four weeks," and they can get it.
Yeah.
There's not a lot of incentive for them to build inventory, you know, particularly when we're in a not a great economic situation right now, and there's some uncertainty. So with that, I think they're being conservative. And until they get burned and place an order and they can't get it fast enough, there's not a lot of incentive to build out. And we've got a lot of inventory, as I say. But, you know, at some point in time, that dynamic will change.
Mm-hmm.
Inventory will need to increase 'cause we're on the lower ends of what we've seen historically in the channel.
So let's talk about, and we talked about your, the inventory position of your customers, talked about the channel and demand, right? So ISM in the U.S. still contractionary, auto sector data not super, not super great for the next 12 months or at least forecasted. Could you talk about, any varying views from just the broad health of those end markets? And then where can you guys see secular growth that gets you a little bit better company growth versus those end markets, which are.
So.
Trending kind of flash?
You know, in terms of secular growth, where we're seeing good growth is back in data center products, obviously, right. So our data center product groups are growing, new designs, new activity. Second market is in communication market, you know, whether you know that's the Ciscos, the Nokia, the Sierra, you know, those are all growing i n the A&D markets, right? Those are all growing markets for us.
Mm-hmm.
So within those three, even though ISM may be down, those three markets are all growing, real demand, right? And so those are doing very well for us.
Have you guys broken out as a percentage of total revenue how big data center is today?
So we, we've broken out data center and compute combined.
Okay.
When that was about 19% of our revenue l ast year. W e haven't given any more granularity than that, but it's a combination of both.
I'd imagine growing as time goes on.
Yeah, and A&D, we broke out as about 18% of revenue, last year.
Yeah. I wanted to jump into that next. So could you talk about A&D? I mean, in terms of the companies that we're seeing here, the verticals that are growing outside of AI, you can't really name a better one other than A&D. You have obviously ground-based radar arrays, satellites to some certain extent.
Yeah. Hypersonics.
Where do you guys play?
Yeah. Yeah. So we play in all of them.
Yeah.
Quite literally, right?
Yeah.
Whether it's hypersonics, whether it's Patriot, whether it's ground defense systems, drone defense systems, those are all Microchip marketplaces. There is nothing that leaves Earth orbit without Microchip. There's no defense program that doesn't use our devices. You know, the Mars Rover had over 130 products on it. The James Webb Telescope that transmits those beautiful pictures back to Earth uses our drivers and, and various technologies to, to take those pictures and, and send them back.
Mm-hmm.
Whether it's the DART to move, or Artemis, all our devices. And then, you know, we recently, in fact, we just got it out of the wafer fab. We were contracted by NASA to develop the next-generation space computer, which we're calling a high-performance space computer. It's an octal RISC-V 64-bit machine. And so the space community has essentially been using an Intel 286 for forever. And now, we have given development systems to almost anyone doing anything with aircraft and space, whether it's Boeing or Raytheon or Thales or Safran. They're all now developing code for that next generation of space computer. And we'll probably be sampling that mid this year to those customers, maybe even earlier.
So, it sounds like more broadly inventory coming down in these broader end markets. But if you look at wireline, data center, A&D, like, these are areas of secular growth where don't just look at these end markets and what they're doing. Look at where the secular growth drivers are that are becoming an increasing percentage of revenue. That can drive some more top-line growth.
Yeah. So if you're an industrial customer and you're doing Industry 4.0 and you want to have vision and sensors or robotics, you know, RS-232, 485, CAN, LIN, or out of speed, right?
Mm-hmm.
You need to upgrade to Ethernet, and a lot of these industrial accounts are moving wholesale to wired twisted pair Ethernet. We're seeing a great deal of design activity, and those will be layering into our revenue in 2027 and 2028.
So if you look at the different analog and mixed signal stories kind of across the space, there are various ownerships of own fabrication facilities. There's the Fab-Lite model. There's the full ownership. You've seen TI go in one direction. Remind us where your strategy is there. And then as inventory comes down, could you help us think about the leverage that you get on the operating model as you start to see revenue growth with these opportunities?
Yeah. So we do about 37% of our wafer fab in-house. Today that's in two factories, one in Oregon and one in Colorado. We in this last year shut down our factory in Arizona as we rightsized manufacturing when Steve Sanghi came back to the company and we're focused on inventory reduction. But we're in a good spot. We've got lots of clean room space to grow into. In those factories, we actually have about $400 million of equipment.
Mm-hmm.
That is sitting undepreciated, on the balance sheet and ready to deploy as it needs it. We had about a $50 million underutilization charge running through cost of sales last quarter. That will come down gradually over time as we grow back into our capacity. But we're reducing inventory right now. So we are producing at a level that is substantially less than what we're shipping at. And then we'll have to pull the levers as we go through to increase capacity, at the right point in time. We actually are producing more out of our wafer fabs this quarter than last quarter. And I expect that to continue as we move through 2026. So that is one lever on gross margin. The bigger lever on gross margin in the short term is we have been taking accounting charges for what we call inventory reserves.
Mm-hmm.
Writing down inventory based on our policy. Last quarter, that was almost $72 million. I would say a normalized level for that 'cause there's always some level of inventory reserve is between $15 million and $25 million. We're sizing this as about $50 million in quarterly benefit to gross margin that is going to come back to the P&L sometime over the next three or four quarters.
Mm-hmm.
So, even when we initially guided this quarter down 1%, now we're guiding up about 1%. You know, gross margin was improving because these reserve charges are going down. And it's driven by two factors. One, inventory on the balance sheet is coming down. And the basis of the calculation is looking back on a trailing 12-month revenue, which has been declining, and that is inflecting upward at this point in time. So that's a big driver. You know, that is going to benefit gross margin, and it's pretty material. So we've got a long-term target of 65%. Last quarter, we were 56.7%. We have a very clear path to get there. And then on the operating expense side, our target model is 25%. We're more like in the 32%+ range today. So we've got a long ways to go on that.
That's gonna be driven by revenue growth over time.
Yeah.
Us being cautious and appropriate as we make investments in the business, but OpEx dollars will be increasing, but OpEx as a percentage of sales will gradually come down.
Helpful. I want us to pivot a little bit to the geography conversation. So something that we've seen in microcontrollers, and this has been a worry a long time. This is nothing new. You guys have heard this, is just on at least the low-end side, 8-bit microcontrollers, you're seeing an increasing presence from China. Is that in your eyes completely domestic China? Are they able to sell that product globally? Are you seeing more competitive dynamics with China in the markets that you're playing in today?
You know, I a lot of that's just local.
Yes.
Right? And a lot of that's 32-bit. And you know, most companies, you know, for security reasons are not building in programmable devices outside of that, you know, especially for infrastructure or tax dollars are in play. So that's a very different thing there.
Mm-hmm.
You know, so a lot of those products are available in China. We, if you look at where Microchip has been going, you know, we've been moving much more to the higher-end products. The bigger story with Microchip now is really connectivity, networking, compute, you know, those types of devices in terms of where our growth is. I mean, we're still investing quite a bit in microcontrollers. We've done a lot of work in consolidating, improving the tool suite. We announced an AI code assistant that goes with all of our microcontrollers where we've grounded a large language model and then cuts development time on our microcontrollers anywhere from 40%- 60%. So we're helping customers to be much more productive using the devices, right?
So it's not always the absolute cost of the device. It's the software suite and the tools that support getting that product to market. I think Microchip still leads on that tools environment on microcontrollers.
Yeah. So you're saying essentially the slew of investment in semiconductor equipment that's gone to China. We've seen it. We've historically played in an area where there's competition. We're moving away from that. There's not some sort of cliff that we're approaching here.
No.
Where some part of revenue that we're addressing today needs to move away over time with China. It's really we're moving on to a new spot.
Yeah. You know, we disclose what our China revenue is in our public filings. It's about 18% of revenue.
Mm-hmm.
We believe about half of that goes to multinationals and manufacturers and free trade zones. Comes back to the U.S. and Europe. So we've got 9% exposure roughly to domestic China. And half of that is in products that China competition can't touch.
Yeah.
So we're talking about 4% or 5% of revenue exposure. And these customers chose Microchip 'cause we had the right product, right? A lot of the applications that we sell into have very long lives too. So, you know, there's some pressure there over time, but I don't think it's material to the overall business.
You guys have the benefit of conversations with customers that we don't. We look at the ISM data. We look at the auto data. What should we be focused on when we're looking at the health and well-being of these markets broadly? 'Cause you've been doing a lot of this in the face of poorer end demand still, right? Working through inventory.
Right.
Cleaning up the channel. Like, this has all been in an environment where you really haven't seen what is normalized auto and industrial growth. So what are you looking for as the key sign for markets turning around? I think the market feels like there's been a couple false starts here. What should we be paying attention to?
I, you know, I think as interest rates come down, overall markets start to improve a little bit more. I think, as U.S. savings rates improve and consumer confidence gets in gear, overall markets start to improve. I mean, we got to a point where U.S. savings rates were as low as 3.2%. And that is just an unhealthy spot to be at, right? And so now it's come up much, it's improved greatly. And so, you know, hopefully we start to see some of that turnaround. When it comes to manufacturing index, you know, we're starting to see a lot of industrial customers moving on to new designs. So you've got the new security standards, CRA standards coming out. Entire product lines have to be upgraded to meet these new security standards.
And so, what we're seeing is a lot of older designs need to be redone, right? Upgraded on security, upgraded to meet new energy standards. And so we're seeing a lot of refresh taking place within those particular marketplaces.
You've talked about improving through cash flow and dropping leverage. What are your actions that you're gonna take in the near term in terms of capital allocation, that get you there?
So we are happy to say that, you know, this quarter, our free cash flow covers the dividend. And that has not been the case for the last, I'll call it, several quarters. So, you know, that, that's a good turning point for us. And that's just gonna accelerate as we get into 2026. So we are fully committed to the dividend at the levels that it's at. Investors should not expect an increase in the dividend in the short term. And we also are not gonna be in the market buying back stock because we need to get the balance sheet healthy and bring leverage down. But, I think those metrics are going to improve greatly as we move through 2026. EBITDA is gonna be growing. The debt, overall total debt on the balance sheet is gonna be coming down. And, that's the priority right now.
Get the balance sheet healthy, be committed to the dividend, but not be in the market buying back stock until leverage gets down to a much more reasonable level.
Rich and Eric, it sounds like things are at a turning point here. Let's hope that this is, the jumping point we've all been waiting for the last couple of years. Thank you so much for being here. I really appreciate it.
Okay. Yeah. Thanks for having us.
Okay.
Thanks, everybody.
Thanks.
Thank you. Thank you. Appreciate it.