I guess we can get started if we get any stragglers. Is it still morning? I guess it's sort of morning. Yeah, good late morning, everyone. I'm Stacy Rasgon. I cover—sorry—I cover the U.S. semiconductor sector here at Bernstein, and it's my great honor to introduce our guest today, the CFO of Analog Devices, Mr. Rich Puccio. Before I start, I want to mention if you have questions you'd like to ask during the presentation or have asked, you have a link to our pigeonhole forum where you can submit them, and we will have time for Q&A at the end. Now, of the companies that I cover, Analog Devices is probably, if not the most respected, an extremely high-quality franchise in the analog space, particularly as it relates to signal conversion and processing. And they've really transformed over the last—so it's probably 15 years now.
I mean, they rationalized their product portfolio way back in the days, focusing on high-value applications like whatever the market. They have executed on a really well-done hybrid manufacturing strategy that has offered flexibility as well as supported margins in downturns. They have not been shy to take advantage of inorganic opportunities where available to boost the franchising growth potential, and they are now embarking on more aggressive capital return than we used to see a few years ago, which is good. To tell us all about it, it gives me great pleasure to introduce Rich. Thank you so much for being here today.
Thank you, Stacy.
Yeah, there we go. You were here last year or two, and I think you just said that was your sort of very first sort of public appearance as CFO of Analog Devices. You have been in the job now about a year. I'm just curious, what have you, what surprised you? Any regrets?
Certainly no regrets. I think when I sat here a year ago, I talked a little bit about one of the surprises to me, even though I did a ton of research before I joined the company. One of the surprises was digging in to see how sticky and how long the products are. Stacy talked about the analog franchise. When you get inside and look and see that we have products that are still producing revenue 25 years after they launched.
Do you still show that chart?
The sediment chart, I think, is what we've gotten to calling that, is showed. The thing that became even clearer over the last 12 to 15 months as we continued through what was, for us, the second worst downturn we've ever experienced as a company was the resilience that our business model demonstrated. A big piece of that is what you just described as our hybrid manufacturing process, which gives us a level of flexibility as we go up and down in the cycle to manage the utilization of our factories. At the same time, the other thing that was happening, and I think this also speaks to our resiliency, is our sales teams and field teams were cranking on generating new design wins across our portfolio. If you go back to the prior years where we were generating peak amounts of revenue.
That was two straight years of very strong double-digit growth in design wins. For us, that is a really good indicator of the potential for growth in the three to five-year out period. Seeing that resiliency in the face of what was very challenging, I think Vince has been quoted a number of times saying it is the worst inventory correction he has seen in his entire career. For the company to manage through that and hold the line pretty well, considering the severity from a margin perspective, I thought that was an incredible testament to the business model.
Maybe it's worth jumping ahead then to that margin question. I remember when I first started this job, I think the peak margins of the business were maybe upper 50s. I mean, there were some actions taken around the financial crisis. There were some business exits. I think you're one of the few guys who actually get paid to sell a cellular modem business. I don't think anybody else has managed it since. There were some other business exits. There were some manufacturing footprint rationalization back in the days. There was Ireland and Santa Clara, I think, and Cambridge in Massachusetts. I remember we came out of the financial crisis, and all of a sudden, now the trough margins were higher than the peak margins used to be. They were kind of like high 50s to low 60s.
There was sort of further improvement. You did some fairly accretive acquisitions. You bought Linear, which I think had the highest gross margins, like bar none in the space, like mid to upper 70%, and then Maxim, which had 70%. You sort of talked about gross margins probably holding in that 70% range. It did not quite get there, but high 60s for the '67, I think is where it bottomed maybe-ish.
67 and change.
For the magnitude of the revenue decline that we had around the cycle, I think is pretty damn impressive personally. Maybe you could talk a little bit about that manufacturing, the hybrid manufacturing strategy. How does that work? My understanding is you're licensing processes from TSMC, and the capacity is fungible. How does that actually work? Because there aren't a lot of other companies that have been able to pull something like that off at scale.
Yeah, so great points on the margin. Yes, we were disappointed. In defense of my IRR team that built the long-term model in 2022, I think they contemplated some downsides. I do not think they contemplated the severity. For us to hold pretty close to 70, we felt really.
Pretty close.
Really pretty good about that. What the company has done, what we've done from a flexibility perspective is what you described. We have the ability to produce at this point, the majority of our products can be produced in our own factories or in our external partners.
What percentage of products and of revenue?
By the end of this year, we expect about 70% of our products to be able to be manufactured internally or externally. I always caveat this. That's not a statement that we would move 70% of our stuff out. We still tend to run about 50/50 internal, external for.
Is that 70% of revenue as well, or?
Yeah, it's 70% of revenue as well. If you think about that, what it means for us is when things were going crazy in the upturn coming out of the pandemic, we were able to go external and get extra capacity from our foundry partners. That helped us mitigate our factories during the peak of the comeback. We were running way above normalized utilization rates. I think even getting to 100-plus percent at one point, theoretically, but significantly above the sort of ideal utilization rates. Look at the peak to trough drop across any of our markets.
How much was it? It was more than 30%, right?
I think it was about 31, or did it go to 34 in the last quarter?
It was north of 30%, peak to trough. For us to be able to maintain a margin in the high 60%, part of that was our ability to bring capacity back in. You can't do it overnight, but with some notice period, we can bring back some of the products that we might be using at an outside foundry into our factories to help keep the utilizations from going down. You talked about this before, right? If you look at the peak margins that you were talking about, say you go back even the most recent peak margins in the 2022, 2023 time period, where we got operating margins up in the 50%. If you look at the revenue drop from then, there were sort of two primary drivers to the margin hit.
One was underutilization in the factories and the variances that got thrown off was a very significant headwind for us because we were well below utilization levels that we would be at historically. The other piece of that was at the same time, the industrial was feeling it pretty hard. If you go back and look, I think in 2023, our industrial business was about 53% of the overall business. It's also our highest margin business. Now you go back and look at the last quarter, it's down to 44%. Those two items were pretty significant downward pressure on the margin. To hold other than, I think, one quarter, we did dip below, to hold greater than 40% operating margin at the bottom of the trough. That is attributable to that factor.
Now, the other piece that, because it'll come up eventually when you talk about OpEx, the other thing is we have a variable comp plan. It's in our proxy, so I'm not disclosing any state secrets. Our variable comp plan is driven by operating margin %.
Revenue growth.
And revenue growth. In the trough, there is zero revenue growth. At 40-ish operating margin, there is not a lot of variable comp. We are now in a period where we are moving up a little bit in the operating margin, but significant year-over-year growth. You saw it actually because it was in, we just talked about it in my prepared remarks. The impact of variable comp was the entire growth in our sequential growth in our OpEx. We have been talking about this since I think we did guidance in Q4. For fiscal 2025, we expected to see less operating leverage than you would normally see in a year with this much revenue growth because we are going from paying very low variable compensation to a pretty significant step up.
If you think and you move out a little bit longer term or even a year later, the step up in variable comp, even if we continue to grow in 2026, does not look like it did in the current year. We would expect the operating margin expansion to improve as we get into the outer years. If you go back to your original question, that ability to swing because we have qualified those products gives us an advantage over many of our competitors.
That must be difficult to manage. You have, what, 100,000 different products?
I think the last count's 125,000. In fairness, a much—sorry? What did I say?
Customers, okay.
70,000 products.
75,000 SKUs. I got it backwards. Anyway, if you break it down, a much smaller percentage of the 75,000 SKUs generate the majority of the revenue. Yes, it is. At any given point, there is a subset of our SKUs that are not moving. The fast movers make up a bigger piece of that. It does make it a little bit. Yes, kudos to our operations team for their ability to manage that with our external foundries and our internal factories.
Got it. I do want to talk about growth. You guys have had, maybe long-term is not quite the right, but you give typically three-year targets for growth. It is 7%-10% at this point off of, I cannot remember what the base was. Was it base 2023? I cannot remember. In 2022. Okay. We are almost through it. Maybe you could give us a little bit of color on where that growth actually comes from? The second question is, you said something on the earnings call, which I think was a statement for this year, but you felt confidence in that target. I mean, what are the drivers behind those targets?
Sure. Actually, I'll do a little bit of a recap of how we actually built out the model when they're doing it. Because I spent a fair amount of time in my first month at the company trying to make sure I was comfortable with the long-term models. If you think about the historical growth in the space and say it's a 5% kind of CAGR business. Then we looked at it and said, "All right, what are the things that differentiate us, whether it's from an idiosyncratic driver or other things that will allow us to grow it faster than that sort of CAGR that we've seen historically?" One, the pricing dynamic for ADI has changed. If you think about the historical Moore's Law world, where every year SEMIs went back and gave back big slugs of their savings to customers, that changed, right?
Things became very inflationary. That model did not continue to work. We have been pretty disciplined. One of the things we have not been doing has been giving back large slugs of the price increase. If you go back again, 2022, when we did our analyst day, we had a pretty significant portion of our growth, and I think we talked about it, was driven by price increase. We saw our pricing tailwind in 2023. As I have talked about, 2024 was pretty stable pricing for us. We expect 2025 to be pretty stable. If you just think about not giving back that pricing gain, we think that is worth at least a point of growth off of that 5% baseline. The second piece we have talked about is when we acquired Maxim, we set a target out there to get $1 billion of revenue synergies by 2027.
Okay. If you think about the progression we're on in 2024, we did sort of the tens of millions of synergy revenue. In 2025, we'll do hundreds of millions, and we are on track to do a billion dollars' worth of revenue synergies for Maxim. If you go off of a $10 billion base, that's another point. The third point is the industry tailwinds where we participate across a number of the sort of current biggest trends, whether it's automation, electrification, robotics. You think about all these tailwinds. The other piece is the idiosyncratic parts of the business where we've got specific design wins in our end markets that are now producing revenue. If you think.
Like what?
If you think about, and we've talked a lot about this this year, ATE, our test business, has been growing gangbusters. That's being driven by the, as are many of these, being driven by the AI boom, right? You think about the high-performance compute, the high bandwidth memory, those all require more testers. That's been a big benefit for us. That has been growing significantly for us. Also on the AI side, if you think about what's going on from a power and connectivity perspective, we have wins in both of those areas that show up in our wireline/data center business. That's strong growth. In our auto end market, we've got GMSL, which is a fantastic technology.
What is that for the audience here?
It's the connectivity for video inside of an automobile. For us, it's a chip at the edge, which is at the microphone or the camera, and, excuse me, at the camera, and then back at the central compute. That GMSL has been powerful. That's actually one of our big synergy drivers, GMSL. If you look at aerospace and defense, we've got a number of design wins in there. That's a market we expect to continue to grow given defense budgets everywhere in the world are being increased. We think that's a good piece of a tailwind for us. The other piece is, over the last number of years, we've significantly diversified our consumer business.
We have a number, we've talked about this, a number of new design wins that are ramping, and we're seeing growth across the portfolio, whether it's in handsets, whether it's in hearables, wearables, gaming. Those are all areas continuing to grow. I think we've grown where you'd see that in our consumer business. We've grown consumer, I think, 30% in three straight quarters. Feel really good about those. Those from an aggregate, just in 2025, we think those are $500 million worth of incremental revenue. We think in total, the idiosyncratic and the market tailwinds are worth another $1 billion. That gets you sort of above the midpoint, a little in the seven to 10 above the midpoint. We think the additional areas just is better execution across the board in those areas can get us to the high end.
Got it.
Like we said, we feel pretty good about where we're going. I did make a comment on the call.
I wanted to ask about that.
Are you going to ask?
Yeah, because.
All right, I'll let you ask then.
The question is, you said you felt for the year, you felt increased confidence about hitting the high end, which is 10%. I mean, that would actually put Q4 down a ton. Did you just kind of, did you misspeak there or what?
In hindsight, I would call it a misspeak at the time. I'll tell you what I was really just trying to do was express that I had more confidence in the year 2025 than I did 90 days ago. Because if you just take a normal seasonal quarter for us in the fourth quarter, we will be well above 10%. More like 12% if we can.
Unfortunately, you know us. We do.
You all went. He just soft-guided the fourth quarter down 7%, which is not what I intended to do. In fact, the seasonal quarter would put us above the range. That was a confidence statement, not a guide.
Got it. Okay.
Sorry for the confusion for anybody who heard that comment, but that was just me being more confident. The confidence stems from we've started saying in Q2 last year that we had hit the bottom and we were going to start seeing growth. We saw sequential growth in Q3. In Q4, we were almost flat in Q1, but a slight decrease in Q1. One of the things that happened in Q1 is what we've been waiting for, which is to see some of the broad market industrial start to grow.
We've been in a downturn there out of, I don't know, eight quarters, 10 quarters for you guys, however long it was.
Yeah, two years. To see some broad market growth to us was one of the first signs that we were starting to see the industrial maybe a start of a cyclical. That trend continued into Q2. We delivered very strong results in Q2 and we're guiding to Q3 with 10% growth. At some level, coming out of that quarter, for us, we were waiting for the because typically you see a bit of an upturn like this when you're starting the cyclical upturn. The question has been for a number of quarters, when will we see a more familiar-looking upturn? Now, do we know the exact slope of that upturn line? No. Does it look like it's getting a little steeper for us right now? Absolutely. That is important.
Tie it back to the first part of this conversation is if industrial starts to outgrow the rest of our business, that's a margin accretive thing for us because at the gross margin line, that's the highest gross margin business we have.
Right. No, no, that makes sense. I guess with that recovery, and it sort of ties to the auto piece as well, right? Auto was pretty strong in the quarter, but you used sort of the dreaded P word on the call. It's funny, like as you sort of look at the stock that evening, it was kind of good until we said, "Pull forward," and then.
I might have gotten a few snap screenshots of the ticker after that session pointing out my commentary. Yeah.
Yeah. I guess.
Thank you, Stacy, for pointing it out again.
Look, but I mean, to your credit, like we were talking earlier, I mean, you call it like you see it and try to be honest with yourselves and with us. I can appreciate that. I guess, what did you see there that suggested it was pull forward? What do you see now? I guess the follow-on is the upside in industrial. How would you know that that is not also pull forward versus just broader cyclical recovery?
Two great questions. We obviously track very closely. We run our whole business on POS, right? How things are selling through, and we look at.
We used to report sell-through, actually, versus sell-to.
We will not comment about what the accounting folks have done to us. We do run the business on the POS signals, but we also, so we are looking at POS bookings pretty regularly. As we watched the quarter progress, we were seeing what we expected as we had forecasted the quarter in growth and bookings across our end markets. What we saw was right around the time of the introduction of the auto tariffs, we saw a spike in the automotive bookings. We also started to see increased turns business in the automotive space. That lasted through Liberation Day and for about another week. After about three weeks of that anomalous behavior, it sort of normalized back down into the weekly ordering patterns that we would have expected.
When we looked at that, first, we did not see that kind of anomaly in any of the other end markets. We do not have perfect visibility that what we said was pull-in was all pull-in. We do not have perfect visibility that we called no pull-ins might have had some. We did not see any of the significant anomaly that we saw in automotive. We talked about on the call that we thought the pull-ins could have been half of our beat from an overall perspective. Seeing it normalize back down was one of the signals for us that it was likely pull-in. It was hard for that to just be a coincidence. I understand we had lots of discussion and debate about this, but ADI has, and this predates me, ADI has always been as transparent with the facts as they can.
We continue that trend. Now, I think one of the things we've been super focused on is Q3's guide. I'll just talk about automotive, right? Q3's guide, we guided down, but largely we guided down because of the pull-ins. We don't actually think the pull-ins came from Q3, which is why if you factor out the pull-ins in Q2, we're essentially guiding a seasonal Q3 in auto. Now, what we're focused on, and as you and I were talking before we started, our industry is plagued by the lack of longer-term visibility on where things are headed. We are actively monitoring and tracking what the backlog build looks like in the out quarters because our suspicion is that the pull-in is likely from Q4 or Q1.
We will watch that, which is why we think that the Q3 is today, as we're forecasting it, does not really factor in any incremental pull-in activity. The rules change every day. They may have changed again last night.
I have no idea what the rules are.
Thanks to three federal judges. Right. At this point, we're staying very close to our customers. As a matter of fact, we just had our teams out and met with all of our top 100 customers to talk about how things are going and what the rest of the year looks like. We're not seeing a lot of unusual behavior yet because of the suspension of the tariffs, and we're still in potentially a pause here. That is the wild card here. For us, the impact of tariffs and the thing we worry most about is just the overall macro impact and the potential demand destruction.
Yeah.
Right? You think about the number of places our products go that will wind up getting tariffed if those new tariffs go into place and then the reciprocal tariffs hit. That could cause some demand destruction.
We haven't seen that yet based on the results, but that's the variable we're all planning for.
Yeah. I mean, that's been my bigger worry. It's not so much the risk of direct tariffs on semis. Even if they're 25% away, semi-ASPs over the last five years are up 50% anyways. We can probably handle something like that. Demand destruction, everything else, I don't have any way to get a handle on. It doesn't sound like you guys do either.
No. Look, we're doing all the things that you'd expect. I actually had a conversation with a group of economists at the end of last week. We're paying attention to how they're forecasting GDP and what that might do and how the various economies might be impacted. Unfortunately, at least everything I've seen and heard is that the economy that GDP is likely most affected by this is the U.S., right? China's still forecasting 4%+ GDP growth. We're continuing to watch that. That'll be an important factor for us. Right now, we're most focused on serving customers today, making sure we deliver on our promises because things could change. We got to serve today.
Got it. We'll watch auto into Q4 and Q1 at least.
Yes.
Maybe that's a good segue, actually, into sort of your geographical footprint. I'd love to learn a little bit both about your manufacturing footprint as well as your demand footprint. Maybe if we started manufacturing, you have internal and external. Just how are you organized in terms of where the wafers are actually coming from and packaged?
Sure. I'll do the quick tour. I'll start on the left part of the United States, right? We've got manufacturing capacity out in Oregon and Washington. You come back to the east, we've got manufacturing in Wilmington, Massachusetts, where our corporate headquarters. We also have manufacturing in Chelmsford, Massachusetts, which is largely in our aerospace and defense manufacturing. We have a factory in Limerick. Those are the.
That's Ireland?
In Ireland. Yes, sorry, Limerick, Ireland. Most of our backend is in Asia. Our internal backend capacity is in the Philippines, Thailand, and Malaysia. We have various partners on the front end from a foundry perspective. Obviously, our largest foundry partner is the big one in Taiwan. We have multiple OSATs across the Asia-Pac region. If you think about end assembly and test, it is largely outside the U.S., and wafers are spread across the U.S., Ireland.
Pretty diversified in terms of a global footprint.
Yes. That potentially could be critically important as we move forward depending on how things shift in the macro and the geopolitical.
I guess what about from an in-demand standpoint? Your China revenue, if I was at 20%, something like that, it's lower, I think, than some of your peers.
It is. I think our China revenue is about 20%. We ship in.
By headquarters, yeah.
By headquarters, it's 20. By ship in, it's about 30.
Got it. N o, that makes sense. Do you guys have, there's a lot of companies that talk about a China for China strategy, just give us. Do you have anything like that? Or how do you think about that?
I think I might, maybe I said this. If I didn't, we have design and sales teams in China, and then we have some manufacturing capacity in China. The majority of our stuff is not manufactured in China, but we do have China manufacturing capabilities.
Got it.
With a partner, not our manufacturing.
Okay. So do you have local foundry partners in China?
Yes.
Got it. What are your broad thoughts just on the current China demand situation? Because it seems like we've seen, I think, auto strength has been pretty good in China, not so strong in auto elsewhere. Industrial maybe is hit or miss, although maybe that's starting to come back. What are your broad thoughts on that? Have you seen any impact, just given everything that's going on, any impacts in China?
Interesting. China was the first to go into the downturn for us. It's been the first to come out. We have posted very strong results in China for three straight quarters. A big chunk of that has been the auto demand in China. One of the interesting things in which I think is an important part of this recovery is we saw pickup in all of the end markets in China in the second quarter. When we look at that, that is a big change because we were still at least 50% off our highs in all of the end markets in China, except for auto. Now starting to see some growth in the broader market. Our footprint in China is very similar to our global footprint. Seeing growth in the industrial in China is a good strong signal that we're.
Still more off the highs, though.
Still off the highs, but a good signal that we're starting that cyclical upturn that we talked about. The other thing is, when we looked at it, if you just look at the broad macro demand, we exited the quarter with book to bills above one in all the geographies and essentially almost all of our end markets. I think we were slightly below on auto, but almost at parity. Really strong exiting the quarter from a demand perspective in what we're seeing.
Got it. Your larger peer in China has been rumored to be being very aggressive on price. I think there's a view that ADI is maybe less exposed to pricing and competitive pressures, particularly in China, given the nature of the portfolio and the value out of the product. Would you agree with that statement? I guess, what have you been seeing regarding the competitive environment, both with local competitors as well as the multinationals in China?
I would say from a pricing perspective, China is clearly the most competitive place. We are not exempted from that competition. I do think that what gives us the ability to withstand some of that competition is, one, we tend to play more at the higher end of the spectrum from a performance and value perspective. We tend to try to be first into a solution where we can leverage our domain expertise. On top of that, we add our local knowledge and our analog knowledge, which we have a significant amount of history on, to be able to capture that high end. The other thing that we are doing is getting more of the system value or driving system value. That is a place where we play a ton.
What does that mean? Is that software? What is that?
It is software. It's also the ability to help integrate into various systems and allow them to perform better because of the, Jeff likes this word, the elegance of our solutions. The other piece of it is, I guess, two other pieces to back to your how we're protected on this is, one, our footprint there is very similar to globally. It is 70%+ is industrial and auto, highly fragmented, very sticky in the industrial in particular, very sticky life cycle products. That is a hard area to attack. Certainly, the design in phase, the price competition is significant. We do continue to win there. We also still continue to capture about 4X the industry ASPs. Now, part of what happens.
What are the industry ASPs now?
Rough number, $0.40.
$0.40. Okay.
The other piece of it is if you go to the other end of the spectrum, a lot of the capacity that's being built both locally in China by the local Chinese companies and also what TI has said they're building over there is much more of the higher volume SKUs. Our game, as we've talked about, has not been to try to pump out as much silicon as we can. It's to go after the value capture. You look at the catalog type parts. That's just not a place that we're equipped to compete with somebody like TI, given their manufacturing model and their efficiency. We don't tend to play a lot in that space. Where you're starting to see the local Chinese competitors pick up is in that space.
We're relentlessly paranoid because they move fast and they're very aggressive, and they will continue to try to work their way up into the higher functions. So our continued drive, and you see this in how we spend our money, is to continue to drive innovation and maintain our leading analog franchise.
How much of what you sell into China is maybe replicable by current Chinese skill sets and capabilities? It sounds like you think that those skill sets are evolving and advancing still. I'm curious what that might look like in five years or even ten years because clearly they're being forced into it, right? The Chinese are putting a lot more resources. They have no choice. They're building a lot of capacity. They're going to have to fill it with something. How do you think longer term? Again, there's a perception at least that ADI may be more insulated from it because of the things.
I think that I always carefully use insulation because I think the threat is real for everybody. I think it is our technological performance advantage and our service advantage that helps protect us. I think that we have a lead of some amount because of our long history and experience. Designing an analog is hard. Just building capacity does not mean you can supply everything that folks need. I think we have to continue to invest to stay ahead of that, or they will continue to take more and more of that market away from us. I think we have a lead. I think it is hard for them to replicate. Where we have seen them take business in China has tended to be in areas where the performance requirements are not as high. If it gets into a it is just okay, then you might pick China.
This applies around the world. If you want the highest performance, you're going to still pick the analog. Even in this super price competition, we are still accelerating growth in our design wins.
It makes sense. I presume even with price competition, you're coming in above, right? I mean, so there's sort of been a bigger push. Vince used to talk about this push more toward value pricing, certainly more versus more than where you were years ago.
Yeah. Like I said, we are super focused on capturing what we get from that innovation and being first.
Got it. Now, I guess the same question on the U.S. side. Again, you've got a competitor that is building an absolute ton of capacity here in the U.S. While that competitor, I believe, actually is actively losing share in China, you can look at it. They've kind of admitted it. They've suggested that in the U.S., though, they think it'll sort of more than make up the difference. How much of a threat is just like they're building, I don't know what it is, six Death Stars or whatever it is in Texas and Utah? How do you think about that? Is there a need over time for ADI to actually be more aggressive about actually being built out here in the U.S.?
I would say in the near term, building incremental fab capacity is not on the table. If you think about what we've done over the last three.
You built up Beaverton a bit.
Well, we essentially will have doubled by the end of this year. We'll have doubled our internal factory capacity from where we were. Given if you do an outlook on where our revenue is headed and from a growth perspective, we feel very comfortable about the factory footprint we have today. Obviously, we only build at a certain lithography in our internal factories. Today, based on the demand we see going forward, we feel very good about where we are from a capacity perspective. The U.S., that I said about China, is going after the hardest problems to solve. One of the things that happened during the pandemic is we got really close to customers. What we find now is they come to us very early in their process and say, "We have this really hard problem.
We need you to solve it." And you know, one of the things we hear pretty consistently from customers is, "You're doing an amazing job. We love all the work you're doing for us. We want you to do more." They are hard, complex problems, and they want to get them solved. That has helped us tremendously get out in front of some of these future design wins because they are bringing us in early because we have my old boss, and I think I said this before, my old boss said, "There is no compression algorithm for experience. We have 60 years of analog experience, so we can solve the hardest problems.
I mean, is that basically the simple answer to the question of how do you compete? Because most of the others, I mean, at least the big U.S. guys would sort of say similar. TI has been around for, I mean, 100 years.
Almost, yeah, 100
What is it that ADI does differently even from the large established guys that would probably try to make the same claim?
I think if you look at them, I think from a relative spend perspective, we invest more of our money in R&D.
What's your R&D percentage?
This year, it was about running about 17% right now. Right? So we tend to run, it's a little bit different. Our first call on capital is R&D versus hardcore capital assets. Now, we had a run here for a couple of years that we've spent a ton on capital. But as we've seen.
This year, you went to what? 8% or 10% around here?
We might have tweaked close to 10% at one point. We have told everybody we expect to get back down into that 4-6% range.
Not 30, right?
No. No, it is not. No. Our team is very prudent in the way they deploy capital. This three-year period was significant for us. Our customers wanted resiliency. At some level, they are willing to pay for that resiliency. We see that. It has been an important thing. It goes back to how much we spend. We have a significantly higher number of engineers than we had three years ago, right? We have more engineers in more disciplines. We have hired lots of software people, lots of digital people. We are ramping up our hiring around AI. We have been hiring for a while because we think that the ability to extract more value out of the core analog franchise comes from having the right digital and software companions.
When you say you're investing in AI, is that investing in AI from a product standpoint or AI from you're using it for yourselves or both?
Both. You think about we're using it at the enterprise level. We're using it at the engineering level from a use of it. We're also deploying it into product, right? If you think about the current big AI build, it's infrastructure and it's data centers. We're benefiting from that. Lots of companies are benefiting.
What do you do there, by the way?
For us, today, our primary exposure is obviously our test business benefits for all the testing around the high performance compute and high bandwidth. If you go into the data centers, they have our hot swap power products in there. We have talked about the new wins where we have got vertical power that is coming online, excuse me, at one of the hyperscalers and it starts shipping for us in the fourth quarter. We are already shipping an optical module that is going to go into one of the large high performance compute companies. Those are the primary areas today in the infrastructure exposure for us. As we think forward to AI at the edge, right? That is the transition from large language models more to small language models out at the edge.
Being able to do the compute closer to where the physical signals meet the digital.
What would you be selling into that?
For instance, think about it. I'll give a couple of examples. One that we're already doing is noise cancellations for hearing aids. That sound is incredibly dynamic. If you have a fixed algorithm in the earpiece, you can't react as well. We are now shipping with dynamic algorithms, mini AI in the—say again? Neural nets. Neural nets. So that you can dynamically react to get better noise cancellation. Another example, if you think about where factory automation is moving and robotics and the increased need for that. Of course, as we build more and more factories, there's not enough people to fill them. You need more robotics. You think about some of the challenges you face with robotics is power consumption, right? It's latency, right? If you've got a robot that's got to react, that's got to—and security.
Having the compute at the edge on the device reduces power consumption, reduces latency, and increases security. The opportunity to build those things into our chips so that in a robot, you can do the compute at the edge where the sensing and the physical data is gathered. I think that is a, and you've heard Vince talk about this, I think that is a really big opportunity for us in the next gen of AI, is getting AI into our products out at the edge edge.
That's interesting. On each earnings call, Vince, he picks a topic and he kind of goes into it a little more depth. I've heard him talking about space and he's talked about robotics and AI and healthcare. What are the areas that you guys, I guess, collectively would be most excited about? Or do I just go back and listen to the last earnings calls that were called out?
I think Vince does a great job calling things out. If you think about some of the broad global trends, our ability to help with the energy transition is super exciting. Think about some of the products we've developed in the automotive space for battery management, grid management type potential, right? The grids around the world are old and in need of update.
Is that the same technology, by the way, the battery management systems, auto that you're doing versus grid?
Same concept.
Yes. It is also we are looking at other ways to help them do more intelligence of the sensing on the grid, right, so that they can better manage what they have. That is a tremendous opportunity. We have talked a lot about robotics, but for us, that continues to be a significant one because of the need for incremental robotics. And two, the change, right? We moved from fixed and now we have the autonomous and now we are going to have the humanoid robot. Every move that way is more content for us. That is super exciting for us as we look forward. One of the parts of the business that has been slow and we have started to see it finally grow is automation. Going forward, again, following those trends of factory automation, we think that is a huge opportunity for us going forward.
Is that sort of where the incremental R&D dollars are going right now?
It is interesting. We spend a significant amount of our R&D just on our core analog portfolio. And then we're looking across what are the most impactful areas and where do we spend our R&D. We're targeting our incremental R&D across these mega trends to make sure we're capturing that going forward.
Got it. What do you guys do on the digital side, if anything? I know you used to have a DSP business. You never talk about it. I do not even know if it is there anymore. Is there anything you do on the digital? I mean, a lot of what you do also be maybe it is not pure analog. Maybe it is more what people would call mixed signal. I do not know.
We are continuing to actually on the digital and software side to develop products to go into or enhance our existing chips. We do have the mixed-signal stuff that you mentioned, but we are also working on more compute type products, I would say.
Got it. Okay. Yeah, you don't talk about it anymore, but it's.
It's harder to move the needle when you get to be over $10 billion, Stacy.
I hear you. Can you talk about the shift? ADI was one of the few companies that actually did attempt to be a consolidator in analog. I think there was a lot of hesitancy to believe that there was value to be added to this idea that we can't get cost synergies because you don't want to fire the engineers. I think you guys have kind of proven over the years that this can work. You've bought, at least in my history, at least three companies that we know of. There was Hitchhike, which was more of a bolt-on at the time, but reasonably sizable and went well. And then.
Amazing Technology actually worked on their IPO.
Yeah. Oh, really? It looks like it's gone very well. Linear, clearly, which was that was sort of the first really big swing. More recently, Maxim. Even the timing on Maxim was really interesting. It was 2020, I think, when they announced it, right?
Yeah.
In hindsight, that was actually good timing. I guess, how was that decision made? Why was the decision made to go out and be a consolidator? I guess the other side of that, why is M&A now no longer needed? Is it just that you can't or that you don't want to anymore?
I'll go back to the beginning. I don't think we've made a decision. I know we haven't made a decision to move away from M&A.
I mean, there was a decision that post-Maxim, and maybe this doesn't hold anymore. I don't know.
Well, no.
They're going to return 100% of cash.
We did say post-Maxim, we would over the long term try to return 100% of our free cash flow. I have talked a little bit about we did not do that last year. That was pretty conscious given all the turmoil. We were trying to make sure we had a little bit of liquidity. We are very happy.