Allegro MicroSystems, Inc. (ALGM) Earnings Call Transcript & Summary

All right. Welcome back, everybody. I'm Joe Moore from Morgan Stanley Semiconductor Research. Very happy to have with us today the management team of Allegro MicroSystems. We have Derek D'Antilio, CFO; and Mark Gill, VP of Corporate Strategy. Thank you guys for coming.

Welcome.

So maybe you could just talk -- start off with a little bit of an overview. You had a CEO change. You've had a good 2025. Can you just talk about what stands out? What's most different about how Allegro is operating versus where you were a year ago?

Sure. We had a strong stretch of years from the time we went public from 2020, we just celebrated our fifth anniversary of being a public company last month, and thank you, Nasdaq for letting us ring the bell again there. So it was exactly 5 years ago in October. We had a very strong stretch with the automotive up cycle coming out of COVID. And automotive has continued to be fairly robust and grow the last several years. But what happened was there was a pretty significant inventory correction in automotive, which is a bit unusual because automakers and Tier 1s don't typically -- Tier 1s, especially don't typically carry a lot of inventory. Well, they did do that starting in 2022 because they were getting incentives from OEMs. Interest rates are the lowest they've been in 100 years. And what happened beginning in the December quarter of 2023, we recognized that there was a lot of double ordering, there was excess inventory. We started to allow customers to cancel orders in that quarter. And we started to see the downturn, particularly in inventory correction much sooner than many of our peers, I believe, and particularly that June of 2024 quarter was quite painful where we came down about 30% sequentially, largely in China where we cleared a lot of inventory. So what are we doing differently? We do have a new CEO that was promoted actually in February of 2024 -- sorry, 2025, just a year ago here. Mike has been with the company for 28 years, and Mike is really all about innovation. He was our CTO prior to this, and he's really focused on innovation and taking a lot of what we're doing in automotive and have done really well in automotive over the last 30 or 40 years and leveraging that to some fast-growing industrial areas that we can talk a little bit about. We've also made some significant changes in our senior leadership team. We have a new SVP of Sales. We have a new SVP of Products. We feel like we're at the end of the late eighth inning of this inventory correction. So the setup coming into 2025 or midway through 2025 here, we feel like we're in a lot better place than we were even a year ago. And there were some lessons learned, I would say, we had in this inventory correction. We have much better models now, internal models for looking at what we'll call our opportunity, our content opportunity, our entitlement models within automotive and particularly with our distributors in the red yellow greens are we overshipping there. So we feel like we're in a much better place. We feel like our customers are in a much better place in terms of how they're behaving with inventory.

All right. Yes. Maybe you could touch on the nearer-term environment. You had guided December quarter to 5% up quarter-on-quarter, better than seasonal. Can you talk about the drivers of that and kind of generally what you're seeing in the market these days?

Sure. So the December quarter, if you look back over the last, say, 15 years, even prior to being public, typically, that December quarter was down 5% on average. It's the only quarter that really has discernible seasonality. Part of the reason for that is we have a very well geographically dispersed revenue mix. So 25% of our revenues in China, 20% is in Japan, 20% Korea, U.S. and Europe make up the rest. And so for example, in that quarter, you have shutdowns in North America and Europe and a lot of the factories, that quarter is typically down 5%. But you're right, we're guiding up 4% sequentially at the midpoint. So we're above seasonal. What's really driving that is continued strength in auto. I think as many people know, auto is turning out to be a better year in '25 than many people thought coming into the year. Coming into the year, it's projected to be down a little bit. Liberation Day came along in the United States and things looked a little bit draconian. Since that point in time, now they're expecting auto production to be up about 3 million units this year. That's number one. Number two, we've seen a real resurgence in our data center business which last quarter was about 8% of our revenue and that's -- those 2 things are really driving this strong December quarter.

Great. And you mentioned this kind of late eighth inning of the inventory burn. You've got revenues that are still significantly below the prior peak. What gives you the confidence that we're that close to the end of this? And where do you think we sit relative to consumption?

Sure. So we look at all of these sort of forward-looking data points, right? And the forward-looking data points are the book-to-bill ratio, the push-pull ratio, how much the pull-in versus the push out, backlog, pricing dynamics. We can also see in the distribution channel, which is half of our sales, we can clearly see on a daily, weekly basis how much inventory they have of our products, what the POS is, what the sell-in is. So we know in September, we under shipped the distribution channel by $5 million. So POS was about $120 million. We undershipped that by about $5 million. That's much better than undershipping by $15 million to $20 million as we had been doing for the previous several quarters. So we're much closer there. I guided for this quarter, I expect distribution inventories to be about flat, a little bit of geography movement there. And then on the auto side, where you don't get that same visibility necessarily because they're using the inventory, we're seeing more in-quarter type of orders. We're seeing more like claims for line potentially lines down in those -- particularly in data center. We're also seeing things that you typically see in an up cycle like our own books having some delinquency on it for parts where we have shortages ourselves.

It's interesting to me when we saw this kind of next period disruption which was a pretty short-term effect, but people went line down pretty quickly. And I feel like there has to be an indication there that there's just not a whole lot of inventory. And like I'm a little surprised given the magnitude of the shortages we saw just a couple of years ago that you alluded to that people would want inventory that lean. Where do you think we are in that? And is there still a memory of that shortage? And can it get triggered by something like that, that sort of changes base?

It's all of the above. I think we're lean on inventories, particularly on the auto side, except for maybe -- at least for us, except for maybe Europe and Japan. U.S. is absolutely lean. China is absolutely lean, Korea is absolutely lean. On the distributor side, we're back to within sort of our weeks on hand of 10 to 12 weeks. It's not lean, but it's back to within that sort of target model for distributors. Will they change their behavior? They certainly remember what happened 3 years ago. However, interest rates are markedly different than they were 3 or 4 years ago. They were getting incentives from the OEMs. So quite frankly, when it's a working capital business, both for the distributors and for the Tier 1s, I'm not sure there's an incentive for them to necessarily go and build a lot of inventory. What we are starting to see, though, is people are willing to pay expedite fees, people willing to buy parts from brokers, right, which are at higher cost than you would buy versus build inventory at least at this point.

Yes. Your pattern in autos has been maybe a little bit different than others where, as you said, you took -- you had kind of a harsher correction and then your recovery has been a little bit more steady and strong coming out of it. Is any particular rationale that any idiosyncratic differences for Allegro that sort of caused it that way or just the way you managed it?

I'll start, Mark, and you can feel free to jump in, but it's really the way we manage it, right? When you look at actual automotive production, again, it's been remarkably stable over the last 5 years, right? It's increased 2 million or 3 million units a year for each of the last 5 years. Only 2 times in the last 3 years as auto production dipped below 10% drop, 2009 and 2020. So it's a pretty stable overall end market. It's really how people manage the inventory reductions and increases back up.

Yes. And I think if you look at that total SAAR number, that's the growth, it's a fairly benign thing. But underneath that, there's some pretty good growth with regards to battery electric vehicles, hybrid electric vehicles, and we are well aligned to that as well. So there's a lot of geographic discussion about what regions those things grow. But at the end of the day, all of that electrification is a good solid growth platform for Allegro.

Maybe you could give us a little color on -- you've talked about your content per internal combustion and per EV. Can you just give us an update there and what the drivers are of that?

Yes. So if you think about the content we have today, you can just take our revenue divided by the number of vehicles, you can get a number and it's approaching sort of like about $9 per vehicle we have today. Our opportunity content in an ICE vehicle is about $40. That's associated with the powertrain, the safety systems, the ADAS systems and those comfort convenience things you have within the vehicle. When we see what's going on in the automotive industry, then there's some pretty seismic shifts, ICE vehicles becoming electric or electrified vehicles. And there's 2 things that are going on in there. You're moving to a platform that's got more onboard chargers, inverters and the likes, that makes a significant difference in the content that we can provide. And that takes us from about a $40 content in a regular ICE vehicle up to about $100 of content opportunity in either a battery electric vehicle or a hybrid. And it's actually a really important point for Allegro, and it's quite different, I think, for us relative to other companies. We don't really care whether it's a battery electric vehicle or a full hybrid vehicle. We have about that same $100 content either of those 2 platforms. So that's the first element. But even then within the ICE vehicles, people still want to make them better. They still want to make those more attractive for consumers. And one of the ways they're doing that is bringing ADAS and safety features into those vehicles. So as we see the human being removed and putting in motors to replace the muscles and the sensors to replace other things, we're moving from hydraulic-based systems into electric-driven based systems, the content for Allegro there approximately doubles from those base systems to more advanced driver assistance systems. So along with the electrification and those increase in ADAS systems, there's really a very strong content. And that's what helps us drive well above that SAAR growth rate.

Great. That's helpful. And I guess what are you seeing in the pace of ADAS innovation? And are you -- does it matter if we're talking about Level 2, Level 3, Level 4? Is there -- I assume there's more content in autonomous, but it seems like you have quite a bit of content in Level 2 as well?

Yes. From Allegro's perspective, it doesn't really matter those higher-end versions of ADAS systems or the Level 3, 4, 5 systems. For us, as soon as you take the driver out of the way, you need -- you need an electric system to be able to drive steering and braking. So it's -- as soon as we move into the content of saying we've got an ADAS, Level 1 plus type systems, you've got Allegro content in steering systems as we move into electric power steering and perhaps even steer-by-wire systems and you have Allegro content in braking systems as we move into the more electromechanical braking and then whatever architectures become after that. So it's basically as soon as you get into ADAS, you get into the domain in which we enjoy.

Great. Thank you. And maybe you could talk a little bit about regional trends, both from the standpoint of the demand picture by region. Is there any difference there? And then specific to China and some of the innovations we're seeing out of China, how well are you positioned?

Yes, I'll start there. So in the September quarter, every region grew for us except for Europe, right? And it's no secret that Europe didn't really grow because -- partially because of seasonality in Europe and partially because it's the area that's probably the most challenged from an automotive standpoint. Now that's 13% of our total business. The U.S. was very strong. Japan came back very strong. China has continued to be very strong. Korea has been very good for us. With specifics to China, and I'll turn it over to Mark in a minute, we have a strong position in China. It's about 27% of our business. About 90% of that is auto. We sell to -- that's a ship to number. We sell to all the foreign companies like VW and Tesla and everybody else who manufacture in China, but also the Chinese global manufacturers, the BYDs, the NIO, Geely, Chery. And remember, all of the growth in China auto projected over the next several years and more likely forever is all export related. So that helps us quite a bit. And in fact, our September quarter, our design wins were led by 2 things: China ADAS applications and data center.

Interesting. Okay. Maybe we pivot to industrial. Can you talk about what you're seeing in the industrial market maybe outside of the data center, where there's some sort of cyclical headwinds still that you're seeing and then some of the bright spots in medical and areas like that?

Yes. So medical is -- so outside of -- so data center, electrification of the grid, trying to make that more robust, for example, the medical business, which was a business that we acquired in through our acquisition of a company called Crocus, those are all solid growth drivers. And each one has got its individual characteristics as to why we see that. And if you think about North America, you got a lot of desire for data center. When you look at the electric grid requirements for that, people are bringing all sorts of technologies in to try to make their local power supply robust, they're local to those things. If you think about medical, it's no -- I think everyone is aware, diabetes is something that is unfortunately prevalent and is growing in our society. As a consequence that the patches that are in that medical product there are -- there's a growth associated with them. And interestingly, as those companies have made those products available and over-the-counter for consumers, it increases the market opportunity for just anybody who wishes to have more understanding about the way that their body reacts to foods and exercise and the likes. So it's another market we're quite excited about.

And then to round out the industrial, Joe, the remaining pieces of our industrial are, we'll call it, broad-based industrial. That part hasn't -- that's been pretty muted for the past year or 2. That's the stuff we sell through the distributors. It ends up in places like Milwaukee tools drills, Xbox controllers, precision sensors. We don't have to do a lot to get that business, but I love it because it's great gross margins because they're buying them in small quantities. And maybe I'll talk about this after, but probably the most exciting part in the future for us in our industrial business is humanoid robotics and there's a lot of opportunity there, right? And that's really taking existing products into, in some cases, existing customers and, of course, new customers there as well.

And who are the customers that you think about in robotics? I was talking to a company yesterday saying there's dozens of them now and you have to have a breadth of distribution-based strategy to serve it as opposed to some of the high-profile stuff that's out there?

Correct. So there's clearly a couple of leading companies in North America who are well known and have some of the most advanced technologies on a worldwide basis. If you think about robotics, there's a wide range of types of robots as well. I think in companies in Korea and Japan might be leading the way in sort of the service and support robotics. You find elderly care, other hospital care, restaurants and the likes of things there. And then clearly, in China, similar to what you see in the electric vehicle market, where there's a point where there are hundreds of companies looking to get into the electric vehicle market, there are hundreds of companies there trying to get into the sort of robotics area as well. And robotics clearly changes from relatively simple devices that you might find in your home through factory robots that are -- let's say, the interesting part for Allegro is joints, right? Things that have motion and sensing associated with them. That little robot around your room cleaning up for you might have, let's say, the equivalent of one joint in it. Those factory robots might have 6, 7, 10 types of joints in them. The reason why people are getting excited about this move to humanoid robotics is just think about your body and the number of joints that you have, and in particular, the number of joints you've got here, right? And we're talking about dozens and dozens of new opportunities in that and so it's a great opportunity for Allegro with motor products that are really the muscles associated with this and then our current sensors and position sensors that are sort of the nerves that think about the positioning, the force, the torque associated with it. So it's a really good opportunity. We're engaged with companies all around the world on doing this, right? This is not just one particular thing in North America that likes all of our companies we're dealing with, and there's great innovations that they're bringing, and we are also supporting them with highly integrated and highly robust precise products for the robotics applications.

And going back to the first question, what we doing differently in 2025. In the past, Allegro has designed all of its products for automotive. We test all of our products for automotive. We sell through distribution to some of these interesting markets originally with solar inverters that have a similar application. So things that have autonomy, electrification, similar to cobots, similar to factory automation. Now we're being much more purposeful in R&D to drive spins and derivatives for the industrial market for these markets, much more purposeful in our sales organization with the new sales organization, and Mark is actually leading our efforts on the industrial side of things by having focused business development teams on data center, on robotics. So we expect that to be a faster-growing area for us, both of those.

Great. Well, in addition to robots, there's a lot of enthusiasm for data center these days in AI as well. For you guys, you had a quarterly record. I think it was 7% of revenue going into data center. Can you talk about the visibility that you have in that business and go through some of your content drivers over time?

Sure. I'll provide some numbers and then Mark can certainly talk about some of the content that's there and some of the data center piece of it. But we had a data center business about 3 or 4 years ago that got to be 7% of our business at its peak. This past quarter, it was 8% of business, right? And the even better news is now it's much more pervasive. Prior to that, it was just cooling fan motor drivers for fans. Now it's fans, it's power management, some of the opportunities they have right there. So that business has come back really fast. So it went through an extended period of inventory exhaustion between the distributors and the fan manufacturers. We're very excited that the business is back, even more excited that it's far broader from a portfolio standpoint and into the power management side of things.

And maybe you could talk about the power management as a lot of discussion of moving from 48 volts to 800 volts, how is Allegro positioned for those transitions?

Yes, I have to say extremely well. So I love the fact that these data center companies and their architectures are tuning into those voltages, which are interestingly and perhaps purposefully the same as we have in the vehicle. So if you think about all of the technology that we've been creating over the last decade have been associated with electrifying the vehicle, much of that technology associated with -- let's start with 800 volts, okay? It's our high-end BEV battery voltage. So the devices that we have there, our current sensors to be able to measure those are supporting well over 800 volt supplies. The products are immediately applicable for those data center architectures there. The high-voltage gate drivers that we are creating and deploying and sampling out for data center customers also designed for supporting those sorts of levels of voltage in the product. So we attach very well to those 800-volt rails and of course, 800 volt is a little high and perhaps a little dangerous at times. So you want to find a nice intermediate voltage and 48 volt is that motor drivers of a variety of our position sensor, our current sensor products, et cetera, and others in our power sensor portfolio. I think one of the things that we talk about inside of our organization is we have something we call true 48-volt products. That is 48-volt is ICE, but you've always got spikes and up and down on these things. You need a product portfolio that is able to survive any of those sort of transients you end up seeing. Our base tech is 110, 120-volt type base technology. So when you've got those 48-volt systems, any perturbations on them, our products are surviving those. That's quite different than a number of our competitors. So we believe we've been investing in this way for the automotive industry for a long time. And as Derek says, it gives us the immediate applicability to go and work with those customers on 48-volt designs or even 800-volt designs in data center as well.

Okay. And you've talked about content in AI servers being a lot higher than traditional servers. I guess where are you in penetrating those opportunities today?

Yes. So as Derek mentioned, we've got a couple of years behind us on the motor drivers that were associated with driving fans. So they were 12-volt single phase, moving into 48-volt, 3-phase fans as those systems become more and more powerful to remove the heat from the systems. One of the things we know about data centers, they consume a lot of power. As a consequence, there's -- we're moving, if you like, those fans, not just from cooling off the compute trays, but now also moving them into the power supply systems and starting to cool those power supply systems. So that's sort of one element of that story. And it's a base layer of the growth we have in our data center business. But AI data centers need more than that and their power supply systems are becoming so -- do they need to be so efficient that the frequency of operation of them is moving from tens of kilohertz to hundreds of kilohertz to megahertz. To be have control of those, you need control loops that are 5x to 10x faster in the components. This is where our 1 megahertz, 5 megahertz, 10 megahertz current sensors, which are unique products where they are using those control to make sure those data center supplies are the most efficient they can be. And then the third layer of this is the isolated gate drivers. Those are just being sampled out to those customers now. Those are, again, trying to make the power supply systems the most efficient. And we have not just an efficiency from a power standpoint, but they're physically smaller than the other products with an integration of the isolated supplies associated with them. So think of the sort of layers of -- we've got the motors. We've had those for some number of years, the fan drivers, and that's continuing, starting pretty much now is the current sensors in our megahertz, 5 megahertz, 10 megahertz products. And then 1 year, 1.5 years from now is where the gate drivers are starting to come in for that. So it's 3 layers staggered out in time.

And these isolated gate drivers drive high power to gallium nitride devices. We don't make the gallium nitride devices. They'll be on the same board. And further than that, we're sampling isolated gate drivers for silicon carbide, think about the EV inverters.

Yes. A lot of innovation in those markets. I'll ask one more question and then open it to the audience. Maybe just talk about your portfolio, your sensor portfolio. I think you released the industry's first 10 megahertz TMR current sensor. Can you just talk about that and kind of where you feel you are with the road map of your sensor products?

Yes, absolutely. So a number -- a couple of years ago, we acquired a company called Crocus Technology. They had some -- what we thought at that point in time was the best in the industry TMR technology. We had some of our own. We assessed ours versus their. We thought it was the right reason to acquire that technology. That is now fully integrated within our organization, and it's fully integrated in all of our sensing portfolios. So it's in current sensors, position sensors, speed in switches and latches, et cetera. Examples, when we first acquired that technology, the primary product there was in that medical patch you described earlier on. It's a switch. It's a very straightforward switch technology using TMR. The next thing we were looking at is a heterogeneous ASIL-D position sensor, which has got a whole sensor and a TMR sensor overlaying each other. They're looking at the same signal. They're making sure that each one is measuring the right thing. That's using an ASIL-D steering system. You just spoke about that 10 megahertz current sensor that is the industry-leading fastest current sensor. So TMR for us is it's a technology, but it can -- we are using it to advance all of the vectors that matter in a sensing portfolio, be that the accuracy, the speed, the power density of these products, the bandwidth of them. And we believe that you've seen some advances from us here, and you're going to see more from us pushing on each of these different vectors as we go through 2026.

Yes. When you look at our road map, Joe, if you look at the products that are continuing to tape out to market this year and next year, a significant portion of those on the magnetic sensing side are now using underlying TMR technology. So that bodes well for 3, 4, 5 years from now, those products being in the market.

Okay. Let me pause here. We have 5 minutes left to see if there's any questions from the audience.

Can you talk about the dollar content that goes into humanoids comparison to EVs or cars in general?

Yes, go ahead, finish.

So in general, the content opportunity on an ICE vehicle is about $40, okay, the content opportunity. On a hybrid electric vehicle or a battery electric vehicle with a standard 400-volt battery, it's about $60 to $65. On an 800-volt battery, battery electric vehicle goes up as high as $100 with those silicon carbide isolated gate drivers.

For robots, humanoids in comparison to EVs, how much -- I mean, you were talking about humanoids, it's more complex, more sensors. What do you expect? I think if you compare EVs, there's only 20 million out there per annum. Robots will be $100 million if you look for the households. So your TAM goes up. What happens basically to your financials [indiscernible]?

Yes. I think the definition of those robots is quite wide with regards to the types of functions and numbers of joints. But if we're talking about those humanoid robots on the right-hand side of that scale, it's more like $100, $110 per robot at this point in time. But I also say that number could increase. The market is very much emerging with regards to products and systems architectures and the technology that we have that would be applicable for them. So at this moment in time, feel free to think of it as about $110. But we'll have an Analyst Day in February, and you may see some updates to those technology numbers at that point in time.

And from an impact on financials today, we're already shipping products into automation, low single-digit million dollars, Cobots already shipping products into that, right? Humanoids are really in the early stages in the sampling design win phase, but you have to be there to really have meaningful revenue in 2030, 2031, 2032. But we do feel like that's going to be a meaningful piece of our business and have a seriously good impact on our financials. A couple of years out, you have to do the right things today from a design win standpoint to participate.

Can we close out with a little bit on -- we got another question.

I just wanted a bit of clarification on your side about magnetic sensing. In particularly in cars, there's a company in Europe, Melexis that does a lot of Hall effect sensors, I think, and they quite a big market share in that. But I see you say that you're the market leader and also that your TMR technology is going to be better or replace Hall effect sensors. Could you explain the competitive dynamics and how that TMR technology help you take market share and where you are in that journey?

Absolutely. So we often use a third-party company called Omdia to try to look at market shares. They position Allegro as the #1 market shareholder with about 23% market share. Second, 4 or 5 points below that Infineon's and third, four or five points below them at Melexis. So that's external reporting. They do that based on their revenue numbers reported. That's just an external view there. TMR technology has the absolute capability to deliver more precision in everything that's measuring. It is fundamentally a faster technology. It's going to deliver higher response rates and higher bandwidth for the products. So when we're looking at position sensors, the accuracy is going to be better. When we're looking at current sensors, the accuracy, the speed of operation is going to be better. And the technology is intrinsically stronger and better, which means you don't need much support circuitry around it, which means it's much smaller. And that gives you very different applications that you could enter into that hall is going to be struggling with. And the example I provided there clearly is if you're in those humanoid robotics and you're looking at the diameter of a finger, which doesn't have a lot of space in it and you need to bring multiple different sensors in there, we believe that the TMR sensing technology is very, very applicable for those small spaces and levels of high integration. It's also very much lower power. Again, TMR technology is what was in that medical patch. So it's very good at extending out lifetime on battery type -- low battery -- low capacity type batteries, not the car stuff, but low-capacity portable type applications.

[indiscernible] cars already...

Yes. So for example, I gave the example earlier on, right, of that heterogeneous redundancy. And see, for Allegro, we have both of those technologies. We can choose which of those is the appropriate one for the particular application that gives the customer most value in what they're looking for.

Okay. Well, that brings us up to the end of our time. Thank you guys very much for being here.

Thank you very much, Joe.

You are most welcome.

Thank you.