Allegro MicroSystems, Inc. ($ALGM)

All right. Good afternoon, everybody. Thank you for coming to the 54th Annual TD Cowen TMT Conference. Really pleased to be joined by Derek and Jalene from Allegro MicroSystems. Both of you, thank you for coming down in the coldest room, I think, on the East Coast. And also, Derek, I appreciate you sticking around New York despite the Knicks being in the finals, not yourself Knicks. Maybe just to start things off, could you both maybe spend a couple of minutes introducing yourselves and introducing the audience to Allegro MicroSystems?

Sure, absolutely. So I'm Derek D'Atillo, the Chief Financial Officer of Allegro MicroSystems -- been here for about 5 years. It's a company that's headquartered actually in Manchester, New Hampshire, and the company was headquartered in Central Massachusetts for actually the last 60 years. After Jalene introduced herself, I'll tell you little bit more about the company.

Jalene Hoover, I head up Investor Relations and Corporate Communications at Allegro. I've been in semiconductors for around 30 years, including SiriusLogix and [indiscernible] Labs, been with Allegro for about 3.5 years.

Great. So I guess between Jalene and I, we have about 60 years of experience in semiconductors. I don't know if that's good or bad. But so Allegro MicroSystems is a company that's actually been around, believe it or not as a company for about 100 years with Sprague Electric founded in 1926 and became Sprague Semiconductor in 1965 doing power semiconductors. Now our company is 60% magnetic sensing. We're the market share leader, global leader in magnetic sensing. We have about 23%, 24% market share, went public 5.5 years ago. 40% of our business is power management. Why give the history is because we have these relationships with customers, particularly in the automotive business that go back 40 and 50 years, having reliability, safety standards met, [indiscernible] certifications in automotive and automotive is 70% of our business. That's exciting, that's growing fast. We'll talk about some of that. 30% of our business are some fast-growing industrial areas like [indiscernible] will talk about like data center and ultimately, robotics.

All right. Thank you for the background. Before we get into the more fun product and content-driven details, maybe we start with -- you just had your earnings call a couple of weeks ago. It seems like things are certainly better cyclically and also secularly for the Analog group as a whole. Could you talk about maybe high-level what trends you're seeing in your key markets? And from a cyclical standpoint, where we are with inventory positioning?

Sure. So we finished our fiscal year at the end of March, we're a March 31 year-end and we finished our fiscal '26 on a strong note. Our sales were almost $900 million, up 23% year-over-year. Our auto business grew 17% year-over-year. More importantly, what we call our focus auto which is xEV, hybrid and EV and ADAS applications grew 30% year-over-year. Industrial business grew 40% year-over-year, which was quadrupling of data centers. So we continue to see strong numbers over the past year. Very importantly, our backlog continues to build. We had a record bookings quarter in the March quarter. So we continue to see restrength across our business in all of these areas. And going back about 1.5 years, Josh, to talk about the sort of cycle dynamics. The automotive industry has actually been quite stable over the last 5 or 6 years. Automotive unit growth has grown about 1 million to 2 million units each of the last 5 years, projected to be kind of flattish to maybe marginally down this year. But if you look at Zoom auto, we're about 40 years, the auto industry production units have only declined 10% in 2 years, 2009 and 2020, right? So it's a pretty stable market. There was an inventory build that we throw like a lot of people a couple of years ago, but we're seeing really strong demand, really strong design wins, and it shifted from being a get bookings to a sort of an operations problem, making sure that we're sensing our customers, making sure that we're meeting our on-time delivery schedule. So really robust demand continues across our markets right now.

And from an inventory standpoint, I mean, I think you guys were more on the proactive side of trying to cut downstreams earlier than others. And you for a while, been flagging that levels are in some spots, quite low. How are you seeing your customers' behavior trending? And any signs of restocking? And overall, I guess, are you comfortable with downstream inventory levels and we can stop asking about inventory?

Sure, so as Josh mentioned, the auto industry went through an interesting time of inventory build that they never did. In '22 and '23, the auto industry, you said, very nice projection of 1 million to 2 million units a year, auto production, good mix shift to XEV over the last several years. But the auto Tier 1 built inventory '22 and '23 that never really happened before. They did that for 2 reasons. One, they were getting dollars from the OEMs to do that in certain countries in Japan and Europe and even in the United States and two, interest rates were the lowest a bit in 100 years right? Three years later, those 2 dynamics are much different. They're not getting those dollars, rates are not particularly high, but they're normal again, right? And so they're not building those. And these automotive Tier 1s are managing their balance sheets and their working capital and so are the distributors. So what I saw was a prolonged period of inventory digestion throughout our fiscal '25 or calendar '24. And what we saw is inventories actually came down below levels that they carried prior to the pandemic. And in fact, customers were placing orders within lead time for a period of time. We could supply those orders a year ago and so could our peers. We told customers we couldn't do it now and now we're ending up having to deliver within 20 weeks on a lead time. So they're dealing with that from across the spectrum. I don't see a massive restocking happening within auto, and I probably understand that, again, because of the working capital dynamics for those companies.

And what do you think is needed to trigger that restocking and getting inventories to levels that you would feel more comfortable with? Is it rates coming down? And I guess on that note, how far is your visibility extending right now as lead times have started to lengthen it?

Yes. Sure. So two things. I think rates absolutely help. At the end of the day, our automotive is consumer-driven good, right? So rates absolutely help there. So rates come down. People need more inclined to continue their build plans. People are more inclined to carry a little bit more inventory. But I don't see them spiking those inventories again to what they did a few years ago when there was sort of the golden screw and then getting money from OEMs. And even if it does happen [indiscernible] to the flip to people's pick up on their revenue, we do want it to be at healthy levels because it helps everyone manage the supply chains -- in terms of the second part of your question, go back to the second part of your question again, Josh.

What triggers restocking? And like are your customers comfortable with where levels are.

Yes. So I don't know if they're comfortable where they are right now, right? Customers are still placing orders not pervasively, but many customers are still placing orders with our lead time, right? And so we're telling them, it's really 20 or 22 weeks -- the good news is our peers saying the same thing that hasn't triggered a restocking yet because they can still buy at brokers at a higher price. I think as that becomes more pervasive it could trigger some restocking. I don't anticipate that being a massive.

Yes. Sorry, it was actually a 3-parter like a good sell-side analyst. It was how far out is visibility extended?

Yes. So visibility is actually really good, right? One of the nice parts about automotive being, give or take, 70% of our business is you get design wins that go out 5 years, right? And so that's really, really helpful. We had record bookings in the March quarter. We have multiple quarters of backlog. That means booked orders. So that's really, really helpful. So on the auto side, it allows you to really plan. I would say that we're starting to get much better visibility from our industrial customers, our distributors, and in particular, our data set of customers to really didn't have a lot of visibility coming into this year and kind of caught us a bit off guard in a good way. We're getting much better visibility there because they're asking us to put capacity in place for us and testing and those sort of things. So I'm getting involved approving testers, approving handlers. But to do that, we need the orders and we need the visibility. So it's much better than it was 12 months ago.

Okay. And then I guess on that note, you and your peers sort of flagged some volatility, specifically in China. That's obviously become literally the most important global automotive market right now. What are you seeing from an order pattern perspective there? And anything we should think about as increasingly China auto mix shifts from local consumption to exports?

Yes. So for us, China is 30% of our ship to revenue, right? And about half of that ends up getting reexported back outside of China, it's the global manufacturers that manufacture in China. It's also the BYD, the GLE, the neos, the [indiscernible] did sell outside of China. And you're absolutely right. All of the production growth auto in China fow now and going forward probably forever will be export-related, all the growth, right? That's good on the margins for companies like Allegro and West supply. We have a great position in China. 90% of our China business is automotive, largely ADAS, largely EV. A lot of it's critical safety applications like electromechanical steering, electromechanical braking, things that have critical safety parameters that while local competitors might be able to beat us on price, they can't meet those standards and those regulations are still pretty strict, right? That said, the export market helps a lot because as they're shipping product into Europe, into the United States, into Japan, that really helps to have Western components in those parts, and that's where us and our European competitors really thrive.

I was just going to say we commented in the last few earnings calls that China focus Auto, which is the XEV and ADAS has actually led those design wins.

I guess on that also, there's been perpetual concern about local competition in China. You guys have seemingly been immune from that. You also, a couple of years ago, went down a path of connecting with some local foundry partners. Can you maybe speak about the local competition, specifically within China? And how beneficial that manufacturing footprint has been?

Sure. And I wouldn't say we're immune to the competition in China, right? China has always been the most competitive market I've ever dealt with, and it always continues to be. The good news is you still win by having competitive specifications ASIL D grade 0 auto safety specifications and the regulations in China for, whether it's emissions or safety or just district is there or any else in the world. So that helps a lot, right? So we went on specifications. We're never going to win on price against Chinese competitors. I don't intend to win on price. In terms of the second part, having a supply chain in China, we're in the process of qualifying a wafer fab in China. We just went live with a turnkey OSAT in China in the fourth quarter, we're already shipping product from an OSAT that does probe, assembly and test for us and ships already from China. We're a fabless company. So we have apps in the United States in Taiwan are our 2 biggest fabs, but we're bringing one up in China. That certainly helps. And I would say that this started 4 years ago, right? It takes many years to bring up a fab -- it has never been a demand. It's really nice to have in China, but it's also nice to have on the wafer side because the cost is better from a wafer. Everything we do is on 200-meter wafers a little bit, we'll call it, legacy technology, 0.18 microns because you have to put high power through these things. And so the preponderance of those fabs are being built in China and the one we use in the United States. So that's been helpful.

Okay. Last one on China auto and then I promise we'll go back to the rest of the world. Your analog semiconductor neighbors in Massachusetts in the recent earnings call called out specifically that I think China or auto orders got a lot stronger at the end of their quarter. That comment surprise you? Is that a track with what you guys have been seeing just direct through the first part of this year?

Yes, it absolutely does. So in our March quarter, seasonally, China auto and China was down. Again, China is 90% auto for us. It was down with a shutdown for Chinese New Year happens every year. But we did see an uptick in revenue in the March, we'll call it, March within that quarter, right? Similar to some of our neighbors in there, there was a pickup after Chinese New Year and Chinese revenue for us, yes.

Okay. And so actually into the content side. At your Analyst Day recently, you highlighted sort of a growth algorithm for how we should bet you guys growing more than SAAR. Can you walk through some of the key sockets that you're most excited about on both sensing and power side? And then maybe help us understand what's tied to ADAS, what side of the EVs, et cetera?

Sure. So we've made a very purpose decision both with our sales team, with our product development team to focus on sort of 2 big markets within auto. Like I said, we've been serving auto for over 45 years, right? And a lot of that was the Western automotive manufacturers, traditional ICE in-cabin safety, in-cabin comfort, LED lighting. We still do that. We still make money doing that. But the focus areas have been ADAS applications, things like electromechanical steering, electromechanical braking where there's a lot of redundancy required for physician sensing, current sensing, motor drivers XEVs, especially when you get to 400-volt batteries and 800-volt batteries with all the DC to DC conversion that happens from 800 to 400 -- all the way down to the 12-volt systems. Any time you have those conversions, we have a lot of current sensors. We have the fastest current senses on the market. And if you look at our auto business now, about 55% of our auto business in total is what we call folks is those 2 areas. Three years ago, it was 35%. So we're making a lot of inroads. That's the fastest-growing area of our automotive. We expect that to grow teams from a TAM standpoint over the next several years. And so we're very excited about that piece of the auto. The other pieces, the ICE piece of it, I would call that sort of a cash cow, where we've had long-time customers. We continue to service those customers, especially the ones that are doing both.

Yes. I think I always assume that more of your content was historically tied to current sensing because of the magnetic sensing exposure. But it actually seems like a lot of that growth is still on the come. Can you maybe speak about where we are in those design cycles with current sensing exposure. And specifically, you bought -- you made an acquisition of Croaks a few years ago specifically to bolster your TMR sensing portfolio. Where are we in the integration of them to your road map and with your customers? .

So let's sort of start with the last piece. We bought this business that does TMI tunnemagnetic resistance. It's kind of the next click of precision. Everything today is on hall-effect technology, everything the industry does largely is on all effect technology. That's been around for 35 years. TMR is much more precise, customers will pay a higher ASP for that in the right applications. The cost is pretty similar to development. So the margins are better general for a current sensor. Some of those products that we've released, like the 10 megahertz current center that's used in data center right now, it's being sampled in data center is a TMR product. We bought that company in to Halloween on October of 2023. And now about 30% of the products in our magnetic sensing business, if you will, are being designed using TMR. So we expect by the end of the decade, that a significant portion of our magnetic sensing revenue will come from TMR. The exciting part for us is customers pay more for the right applications. Not all of it will move to TMR, wherever the value proposition is. But there are about 8 companies we'll call it, in the Western world that does magnetic sensing. I said we have 23%, 24% market share -- within TMR, there's only about 3 or 4 companies that do that. So the pools a lot smaller. We fully expect that our market share should be at least that number, if not significantly higher.

Maybe using that as a segue, so you have the rich legacy magnetic sensing. But you also have, I think, 1/3 of your business is power as ICs as well. That's a much larger but also a much more competitive market. Can you maybe walk through what's your right to win and differentiation in power? Is it the high-frequency switch that you're able to offer? Is it the auto grade qualifications -- what's the strategy and how synergistic is the current sensing portfolio.

Yes. Part of it goes back to the history. We were a power company well before a magnetic sensing company, right? Magnetic sensing we've been doing since the 1990s, but power [indiscernible] since the 1960s, which is kind of interesting. And so a lot of technology there. But the fundamental rights to win all around some of the auto grade technology. It revolves around being able to manage high power, right? So things that require high-power usages in power conversions or step up or step down without losing power, that's really where we shine. . So if you think about 800-volt batteries being in an EV, making sure that you get the most mileage out of that car, the step down of the power throughout the process, all out to the 12-volt systems. That's really where we shine. And the same thing applies to the data center and some of the same things apply to robotics going forward. And then when you layer in some of the newer technologies we've put in power. So really where we shine is we can spin motives very, very efficiently, brushless DC motors, DC to DC conversion, step-ups and step-downs and the most recent product is isolated gate drivers for driving high-power to fast-switching devices whether it's gallium nitrate or silicon carbide.

And I think the isolated gate driver business is perhaps your high single content socket that you have. Is that at the point -- that was through an acquisition a few years ago also is that where it's contributing meaningful revenue yet? Or how should we think about that layering into your model?

It's not. So we bought this technology about 3 years ago. It was a technology at the time, no revenue. Right now, it's in the low single digits. It's in sampling for the gallium nitrate drivers in the data center. It's a very exciting opportunity for us. We took the gentleman who runs our power business and brought our power business to 3% of our business, put them in charge of this isolated gate driver business. It's really a start-up within our business to run it a bit differently, but it's a really exciting opportunity for us. . And the value proposition there really is our isolated gate driver is about 1/3 the size of our competitors. So when you're putting those on a board in a data center where space matters or even in certain parts of the EV space matters a lot, that's important. The power loss matters -- so those kind of things really matter. And we're putting essentially 3 functions into on 1 chip into a monolithic chip in 1 package compared to having multiple chips.

And I think you've highlighted again, but if I'm not mistaken, it's also applicable and works with silicon carbide or silicon high-voltage power as well, correct?

Absolutely. So it's -- we call it a high-voltage power business, right, and does all 3 of those things. The product that we're in sampling right now is with a GaN product that's actually getting design wins. The silicon carbide isolated gate drivers is in development, expected to be released later this year, and the silicon high-power gate drivers is already in the market.

Okay. Let's switch gears to data center, which is the topic of the year, obviously. You're not historically -- you haven't historically been thought of data center story because of all the auto exposure. But you've had a data center business for a while. Can you walk us through -- I think it was 14% of revenue last quarter. What are the key sockets and applications where Allegro MicroSystems is exposed to data center applications?

I'll provide some of the history, and Jalene has been spending a lot of time with our product team, particularly on the data center side, I could articulate some of those things as well. But from a data center business standpoint, Allegro has traditionally sold to the data center largely just the DC fan motor drivers. And that went through distribution. It peaked at about 10% of our revenue, maybe 3.5 years ago. Within that, we were shipping in the inventory. That business went away for about 2 years as the inventory digested. Coming into FY '26, data center was probably 2% of total business. Exiting the year, as Josh said, it was 14% of our business. It quadrupled within FY '26. That part's exciting. A lot of that was the motor drivers. The even more exciting part is now 18% of what we're shipping the data center is current sensors from managing the current that goes into the power racks, replacing resistors. Resistors need things that can dispense heat a lot better, manage power, measure current in real fast time to protect things downstream like PUs and CPUs. That's an exciting part of our business. And then ultimately, these gate drivers again in the data center. If you want to add anything Jalene to that?

Yes. We actually designed in with a few data center customers today with our gate driver solution. And obviously, as you noted, a huge content opportunity going forward. So -- when you look at our data center opportunity today, it's about $150 content growing to $425 -- that growth is driven by the transition and adoption of 48 -- volt and 800-volt technology. as well as continued adoption of our current sensor technology as well as the gauge drivers.

Okay. And so how should we think about like your exposure to air cooling versus liquid cooling? And I guess you mentioned like how big is current sensing today? And like is that the opportunity that could be the largest? Or is it the gate driver side?

So why don't you start with the liquid cooling, Jalene and then we could talk a little bit both current center.

Yes. So the liquid cooling today is what we believe will be an incremental opportunity for us. So that leverages our motor driver technology. And what we've seen this year, for example, is we're actually seeing an expansion of the fan technology from just the racks into the power management. So the data centers have this insatiable need for cooling, right? So there's increased fan opportunity. We've actually -- Mike has brought some toys to several of our investor meetings where we've got the fans there maybe be 2.5-inch diameter for the racks reducing to about half the size for the power management solutions.

And Josh, to answer your second question, it was motor drivers for the past 3 years. This past quarter, current sensors were 18% of our data center business. It was 0 at the beginning of the year, absolutely the fastest part of our data center growth is in current sensors because that's all content gains. That's all replacing another technology that we just have a better product for and there's some competitors, of course, but we're really getting a lot of traction there, especially with our -- we have the fastest megahertz current centers on the market. We have all effect on that does 5 megahertz. We have a TMR 1 that does 10 megahertz. So no one even comes close to that. . And then the isolated gate drivers, like as we said, haven't started shipping, they're in sampling, those have opportunities to be much chunkier, larger sockets themselves. But in the short term, in the medium term, current sensors will be the biggest growth driver within the data center for us first.

Okay. Maybe it doesn't matter because CPU growth is seemingly going to be quite high as well. But can you walk -- is your exposure in data center primarily on accelerated servers in AI racks? Or is it CPU-only racks? Or is it mix between both.

It's both. And really, it's driven by -- it's not tied to any particular architecture, any particular company. higher power equals more legal content for both the fans and for current sensing and for opportunities with isolated gate drivers.

Okay. And your primarily -- like your customer engagements are primarily with OEMs? Like are you at the point where you're engaging with downstream with hyperscale vendors as well.

Generally speaking, our customer or the power module manufacturers, so the light on, the deltas, the advanced energies, similar to the thermal management companies, a lot of them overlap actually. That's our customer. We might go up 1 level in terms of understanding the power after that's coming downstream, but we're not designing ASICs for the GPU CBU manufacturing.

Okay. So moving to broader industrial that through an inventory correction of its own, is that sort of wrapped up and sort of hopefully with the inventory correction behind us for you and your peers, what are the content opportunities within the broader industrial market that you're most excited about?

So I'll touch on the automation robotics opportunity. Those sales, though nation today, they're low single-digit percentage of sales. They actually doubled in fiscal year '26. Most of that revenue today is in more household, think of Roomba as well as Cobot factory automation. But going forward, we actually see that content opportunity to be larger than our automotive. So that content, in fact, I think of the home bought at like a $5 content opportunity increasing to 55 for cobot and about 150 for the humanoid robot. And that's an area where we are focusing our sales team to drive engagement, that's vertical at this time. . Obviously, targeting those companies that we believe will be the highest runners. But this will be an evolution as we go through this process. So it's -- right now, it's obviously the engagement they're going through prototyping next phase is to develop solutions that will probably be in the hundreds to tens of thousands of units ultimately to hundreds of thousands and then millions of units. So we talked about on our last earnings call that we had a win in China, a couple of wins that we called out, one of which was a 90 IC unit content opportunity. And that's important because the number of ICs in these wins is significant. And we have a technology today. It leverages our sensor technology our position centers, our current centers, our drivers. And we do believe that our TMR sensors, in particular, would be of greater value, not only because of the level of precision involved, but they're high efficiency in the really small form factor, 1 millimeter square die. So when you're thinking about our longer-term vision of really owning the robotic hand and you've got really small joints, there's a lot of them, and you're packing a lot of content into each of those joints. That's an opportunity we're really excited about.

And this is another area that we really leverage our auto-grade relationships, our auto-grade technology. When you think about a lot of these companies, at least initially, that are doing production level robotics, right, take China out of the equation for a second, it's the Hyundai Mobis, is through Boston Dynamics. It's the Tesla the Toyotas, right? It's these companies we already have relationships with the area of the reliability data. And of course, there's a whole host of companies in the West Coast that are non-related to auto that we're already engaged with in the same thing in China.

Okay. And any time line at which you would expect robotics to be a material portion of revenue? I know it gets talked about a lot by the whole industry. Nobody ever gives me any numbers on how much it actually contributes. So I'm going to try now.

So today, as Jalene said, it's low single-digit millions of our product, right? But it's meaningful because it is factory automation, it is cobot, it is those kind of things, right? Humanoid is minuscule piece of our revenue today. For us, internally, both our business development teams and our product teams -- it's not a question of if. It's just a win, right? Is it 2030 or 2031 -- but it doesn't matter. We have to be there. We have to get the wins today. We have to have those customer engagements today. We have to be ready to capture that with existing products, and that's what's exciting about it. It's existing families in many cases, existing customers, right? So we'll be on those road maps and the products that are being designed today that are go into production today, those won't ultimately be the 50,000 unit ones. These are the beta versions and things like that. I think meaningful starts to get into 2030, 2031. Sounds a long way away. You have to be there today in the design wins. You have to be there today and the samples with the reliability data with the customers to get that opportunity in a few years.

I think to add to that is an accelerator would be a geography like China, where the adoption and advancement of this technology is critical given the population decline. Conversely, I think people need to appreciate that there's a qualification process with advanced humanoid robotics that will take time to go through that evolution. So it's a push-pull factors, if you will.

Okay. I'm going to make a really hard pivot to gross margins now. .

So Derek's favorite type.

Higher is better.

Okay. Let me write that down. So you guys obviously peaked in the high 50s during -- when pricing was peak and everyone is running at over 100% utilization rates. You've walked through some fall-through numbers, but could you maybe spend some time talking about how you expect to go from where you're at today to that mid-50% gross margin target you gave at the Analyst Day?

Sure. So our gross margins peaked for a year at 56% on a non-GAAP basis. Actually 2 quarters, they hit 58%, and that was at sort of the peak of the inventory cycle, pricing, cost structure was different. Over the last couple of years, auto pricing has come back down our commodity pricing is quadrupled, for example, things like gold. And so our gross and then utilization, of course, came way down in the back end. Our gross margins troughed at about 46% on a non-GAAP basis. That was the end of last fiscal year. We exited this fiscal year at 50%, so up over 400 basis points. So that's good. We guided Q1 to 50% to 51%. Our model is to get back to north of 55% and really, there are 3 big pieces of that. One is volume. So volume helps a lot. Our variable contribution margin is above 60%. So putting volume through our back-end facility provides more than $0.60 per dollar of revenue through the model, right? That helps a lot. The second piece really is improving that variable contribution margin. And there are sort of multiple things happened there. One is we're going through a large gold to [indiscernible] a conversion on our wires on the packaging, right? And that's a significant -- it was a 200 basis point headwind in FY '26 alone. So we're going through that process. It takes time to qualify that. It doesn't all happen in one quarter. It will happen over the next year or 2. In addition, there were things like shrinking the die size, which is the biggest piece of our bill of material. Jalene talked about with TMR, they use a significantly smaller die, 1 millimeter square by 1 millimeter squared CFOs love those things because you get a significantly higher yield, more die per wafer. That helps a lot. Mix helps as you start to put more current sensors into the data center. So previously, data center was margin dilutive to our business because it was all motive drivers as we're selling more current sensors, it's much closer to the fleet average. We expect that to exceed the fleet average and move into the back half of this year. And then as we put the isolated gate drivers out there. And then the last piece of it is [indiscernible] factory efficiency improving our overall equipment utilization from the mid-60s to closer to 90%, really getting use out of that back-end facility. So there's multiple levers here sort of all happening at the same time. They're not necessarily linear but I think we have a solid plan to really get back north of that mid-50s on gross margin in the next couple of years.

Okay. We're running out of time. So I'm going to ask my last 2 questions at once. First one, can you speak about what you're seeing in pricing here and now because there's a lot of attention on analog pricing passing on higher input costs and pricing being more firm than it's been the last couple of years. And on that note also, you guys are a bit unique in that you're a fabless company in competing with many IDMs. There's concerns about shortages again. And do you feel like you have enough capacity secured to grow through some of the time lines in which robotics, for instance, might be more meaningful.

Yes. Great question. So on the pricing side, we said it on our call, we're increasing pricing, but it's not across the board, right? We've had these relationships with customers for, I guess said, over 40 years, 70% of the business is auto. So you have to be much more surgical on price increases. In fact, with all the estimates you end up having productivity declines every year. But on our long tail of distribution customers, pricing increases are happening because of input costs, we also are doing surcharges for things like gold, fuel, those kind of things, different flavors of price increases, whether it's expedite fees on shipping and those kind of things. So that's happening. And I think we'll balance that with market share gains. On the second piece, in terms of capacity, we feel like we have the capacity to handle the business we have today and in the near term, in the foreseeable future bringing up these fabs in China particularly for automotive, but also for the robotics, right? And quite frankly, we have 2 fabs in Taiwan. We have 1 fab in the United States that does about 1/3 of our wafers. They're doubling the size of their fab and now for this reason. It's part of an expanding project with some chipsag funding. We have other fabs around the world that we're using. So frankly, for a company our size, we might have too many fabs. That's the bad news. The good news is we'll have the capacity going forward.

Okay. All right. Well, we're out of time. Derek, Jalene, thank you very much for joining us again. .

Thank you very much Josh. Appreciate it. Good luck to the next.