Marvell Technology, Inc. (MRVL) Earnings Call Transcript & Summary

All right. Good. I guess it's still good morning out here in Las Vegas. Good afternoon to everybody else. We'll kick it off with the next company here up on stage with me from Marvell. We have Ashish Saran, who is the Senior Vice President of Investor Relations. I didn't want to miss the S and the VP side of things. So Ashish, thank you for coming and visiting us in Vegas. You guys just reported very, very recently and gave the third quarter guide. There was definitely some moving parts on the supply-demand side of things, more on the supply side in your instance.

But the biggest question I've received post earnings is on your data center segment. It was about flat in the second quarter. People expected a little bit of growth in the third quarter, but you guys guided it down mid-single digit percent. So talk a little bit about the supply versus the demand side and then we'll go into the cloud versus the on-prem side as subsegments within data center.

Yes. First, thanks, Ross, for having us here. So I appreciate that. I think we need to step back a little bit. I mean if you remember, supply has been a challenge for the industry for probably 1.5 years at this point in time. And if you all remember, back late last year is when we set ourselves a fairly, at that time, an audacious goal of driving low 30% year-on-year revenue growth for this fiscal year, the one we're in right now. And really, our view was that supply for us would improve. As we've gotten through, say, roughly 6 months of this fiscal year, I mean, the good news is we're actually significantly ahead of our target. We're now looking at this year, probably ending up more towards the high 30% range. So I think in totality, we've actually gotten more supply than we'd originally planned for. And perhaps we're a bit of a victim of our own success as we've really stretched some of the supply chains, especially on the high end of the product lines. Think of these as typically multilayer substrates, some of the more advanced packaging technologies typically have longer test cycles. And I think that's what we are facing going from Q2 to Q3 is the mix isn't exactly what you were expecting, mostly because of supply. And then clearly, if you look at even the whole year so far, some of our sequential by different -- end markets have really moved up and down, not as an indication of demand, but it's really a reflection of where supply ends up. So Q1, if you remember, we had a very strong data center. We grew it up like in the low teens sequentially, which is fairly unusual, but that was mostly because supply ended up in the right place. So I think most of these movements you're seeing is really supply. Again, as we look forward, some of the supply, which we would have hoped to have come in Q3 is looking more like Q4, which is why we've looked ahead and said, hey, while the mix isn't exactly what you were expecting, Q4 overall, the sequential growth reaccelerates at the company level. And not only that, we expect a lot of that to come from data center, right? As we start to catch up there as well as the automotive market, where our content gains in model year '23 cars start to come into production. So that's quite critical for us to enable that. So overall, I'd say, supply looks a little strange right now. Good news, there are some pockets of supply opening up. Some of the 28-nanometer is one example, some of the BGAs, which are simpler is another example. So I think those are helping us in enterprise networking as we get into later parts of this year into next year. We would anticipate that even on substrates, we should be able to get better supply as time goes on.

And talk about the differences in the growth between the cloud versus on-prem subsegments underneath the covers of data center.

Yes. I think first, the majority of our revenue, certainly within what we report at the data center is from our cloud business. Enterprise on-prem is a relatively smaller business. On enterprise on-prem, these products are things like Fiber Channel, Ethernet NICs, some of our processors. Our assumption long term is that, that's not a fast-going business. That's kind of a GDP-ish kind of business, right, enterprise on-prem versus cloud is really where we expect on average, the end market to grow, let's call it, somewhere in the 20%-plus CapEx range. And then we have our own product cycle. So we expect to do a lot better. So that's kind of the mix there.

And I would assume to the extent you have volatility between the supply or the cloud and on-prem, there's also the supply side of the equation where you're prioritizing that supply given those long-term growth aspirations. So it can still be more of a reflection of supply than demand even on the areas that have been weaker.

Yes, you're absolutely right. I think if you look back, some of these decisions you have to make on where you allocate tight wafer supply, you take those decisions much, much earlier in the year. So some of what you're seeing play out in our different end markets, you can argue the same thing within our auto industrial, but automotive is clearly the faster-growing market. So -- and that's where some of the older nodes have been quite tight. So yes, I think in general, it's fair to assume you typically prioritize -- you need more supply for your faster growth markets. That's clearly a fair assumption. And then when you think of some of our adapter-level products, there is certainly higher complexity. There are a number of additional components, which have also been challenging, right, as we've gone through the year. I think that's a very fair assumption.

So just before we get into some of the other end markets and what you guys are seeing recently, there was some announcements last night or recently from the U.S. government about requiring a license to ship certain things into China. So just under the umbrella of the cloud business that we're talking about, have you guys gotten a letter on that? Does it impact you directly or indirectly?

Yes. We have not been notified, right, which is probably not very surprising given kind of the business we have, which is probably more geared towards optical products and other areas right now. I mean there's obviously worries about ancillary effects, but at this point in time, I don't think anybody really knows. So from a simple perspective, we have not received a notification. For us, it's -- we continue doing our business for now.

Got you. So if we move on beyond the data center side, the second most popular topic that I was asked about after your earnings call was on the carrier segment. And it seemed like it was almost identical in that the good side is good, the bad side or the stable side is worse. And you guided that down mid-single digits as well. Talk a little bit about the wireless versus wired trends underneath that carrier category.

Yes. I think this -- again, I know it's kind of beating the supplier drum, but it's very much an artifact of what's happening on the supply side. So first is our wireless business, which is clearly the focus. Has been growing very consistently, and we guided it up again for Q3. In fact, we hit a pretty important milestone for us, which was we crossed $600 million of annualized revenue run rate in Q2. So wireless is the bigger part of our total carrier business. The wired business is one which is more of what we think of as the steady-as-she-goes business. In Q2, it actually did significantly better than we expected, not because the demand suddenly spiked. It's because we had better supply lined up, right? So while we've had supply issues in other areas, we actually had better supply coming in on wired. And that's why Q2 was higher than expected and Q3 -- it's a lumpy business. It just renormalizes. So I wouldn't read that why demand is down in Q3. I think it's just an artifact of what we did in Q2. And again, I'll go back to if you look at our sequential buy-in market from Q1 to Q3, I mean it looks like it's moving up and down all over the place. But it's really a question of how supply ends up, right? Demand, I would say, in general, outside of consumer HDD has remained quite healthy, right? And I think that's probably the most important thing.

So the last one about the quarter and then we'll get to some longer-term trends. I think the enterprise side is, by my math, has grown sequentially for, I think, 8 straight quarters now. And you guided it up I think about 25% roughly sequentially and 70% year-over-year. I know you'd love to say that's design wins and market share gains, but I would assume there's also some supply allocation going into that equation. But generally speaking, what's driving that sequential growth and probably more importantly, year-over-year growth?

No, it's a great question. I mean enterprise networking is one of those markets, which I think we all kind of take for granted, and it's not an area which investors always focus on. But it's a pretty important market, and this is one where we have been gaining share, right? We had a number of new products introduced. We didn't really see the benefit from those products for a couple of reasons. One is all our customers who introduced their new products, and that really didn't happen until late last year. And even once they did, unfortunately, we were all supply constrained, right? So we really couldn't deliver the share gains we had actually won. On top of that, we've certainly seen a pretty big shift in content upwards. In particular, one example would be an acceleration towards using multi-gigabit Fis versus gigabit and the ASPs are significantly different there. That's a pretty big tailwind. So through Q2, I would say the majority of the growth we've seen has really been a combination of our share gains starting to show up, again, multiplied by higher content. So you actually get a double whammy here, right? I think in Q3, and in fact, till Q2, our most constrained end market, where we were the most delinquent was really enterprise networking. In Q3 is where we start to catch up there, right? So to your point, it's not just share gains and content growth, you're finally seeing some of the units coming through in Q3. And I think from this Q3 point, I would say we would suggest, as you look a little longer term, the growth probably is more in line with our long-term model, right? This is not a long-term fast-growing business, but certainly, we've got our own cycles. Two more things to point out in enterprise networking. As you think into next year, we still have a custom ASIC program, which is still going to ramp, started this year, but it will take till a big part of next year to ramp more fully. So that's still kind of an offset from the market in front of us. And then I still expect more multi-gig Fis next year compared to this year. So we still have a couple of offsets. We're looking forward to next year.

And is that business one that -- when you think about the longer-term growth rate of enterprise networking with your share gains and all those sorts of things. So for the Marvell growth, right, not the market growth, right. What's a reasonable number? I know it's below the 70% year-over-year this year.

Yes. So it will be a little more disappointing than 70% year-on-year. It -- our view is that the end market is probably a GDP kind of market, right? Call it low to mid-single digits. And we generally said we'll do a bit better than that, right? And I think at this point, as we get into next year, we start lapping some of these very high growth rates. I think that's a fair assumption that in a normalized environment -- like let's kind of ignore some of the macro stuff, in a normalized environment, enterprise networking, it's IT spending 1x to 2x GDP. And then we still have, like I said, a couple of product cycles. So we would in a normalized environment, do at least that much or perhaps a bit better.

What are the product cycles that are going to give you the better?

Yes. I think it's multi-gig in particular, right, is where you start to see higher content. And then on the switch side, certainly, a lot more deployment of feature-rich switches in enterprise. As enterprises do their upgrades, I think those are where you get better content. And then like I said, we still have an ASIC program, which is going to ramp over the next year, which is certainly something we have very little revenue. And if you go back a few months, right? So I think those are the 2 big offsets as we look at the next couple of years.

So before we dive into some of the longer-term trends by end market segments, let's just talk about the transformation that Marvell has undergone over the last, say, 5 years. You guys have gone from a heavily consumer-oriented company, storage and networking. Now you have more end market diversity, a bunch of acquisitions, et cetera. The net result is a faster-growing company, clearly. How much do you think the secular growth potential of the company can offset the cyclicality? As investors right now are worried about the cyclical side of the equation, how do you balance the new Marvell's exposure to that versus what investors might have been fearful about considering the history that the company had?

No, I think it's great observation. I mean the whole focus of the last 6 years of transformation has been to really shift towards high-growth markets and more stability, right? And I think if you look back 6 years back, our biggest end market was storage. And within that was HDD exposed to the PC market, but there was no share gain left. So you were basically a run rate business. Whatever the PC market did, would basically translate into your revenue, right? So it was a very much -- a cyclical approach. As we look forward, right, I mean, HDDs is a fractional part of our revenue. It's very, very small. Consumer in total is now basically roughly 10% of our revenue, right? And then the other 90% is really exposed to data infrastructure, right? So if you sit back and think about what are the areas from a secular perspective, you believe the world will need to keep investing in. It's data movement, whether it's within cloud, which is clearly our most secular end market. It's 5G, where you still need a conversion from 4G to 5G over the next few years. And then the other one is automotive, where we're not tied to unit growth at all. We're really tied to Ethernet now becoming the most pervasive technology going forward from a connectivity standpoint. And we've also got a very strong share position there. So I think from a conversion, 90% data infrastructure, probably the best place you can be in semis today, I would argue. And on top of that, I think we've got our own product cycles, which we can certainly discuss in more detail. So from a positioning perspective, I mean, this is what we set out to do a number of years back, and we're at that goal at this point in time.

Any other big pieces of the pie or puzzle that are missing far as M&A, divestitures, acquisitions? Or do you pretty much have the pieces in place to deliver the growth that you're envisioning?

Yes. I think we've done a couple of very important acquisition, which really turned the trajectory of the company, acquiring compute and secured capabilities, we needed Cavium, which also got us on to the 5G cycle. And then we really upped the game with Inphi, which caught us electro-optic we have all the high-speed connectivity links. Marvell on its own already had fantastic technology on the storage side as well as on the switch Inphi side. And then we also made a smaller acquisition, which turns out now to be actually quite important, which is Avera, which is what gave us our custom capabilities. And then Innovium was kind of the last tuck-in, which gave us kind of the last piece of the switch Inphi, which is access to high-speed data center switching. So at this point in time, I'd say we have all the pieces, and we are seeing the benefits now in terms of customer engagement, but they look at us as a full IP provider. And we are now able to organically come up with new opportunities, We've talked about a few of them, whether it's CXL, it's AEC, active electrical cables. So I think we see more potential now for organically driving the next phase of growth. I think we'll always be smart. You always want to keep our eye open and see if there's potential for tuck-ins along the way. But I would say at this point, the big shift is done. I think now it's more about executing on the shift we've done within the company itself.

So let's talk about some of those company-specific drivers of growth. And why don't we start on the data center side? You guys talked about the cloud-optimized solutions, adding, I think, $400 million in each of the next 2 years to your revenue growth. Talk a little bit about what builds up to that $400 million, and then we can -- I'll ask about how sustainable and kind of recession-proof those ramps should be?

Yes. I think this is, again, when we started down this journey post Cavium and we started our first set of conversations with cloud customers, it became very clear to us that what they were really looking for was a partner who could give them optimized solutions, right? And the gamut of those solutions vary from I need a partner. I want to develop the chip myself, which is basically an ASIC, and I need someone to help me do it. which, given the advances Marvell had made in process technology, where we now had a 5-nanometer road map, our SerDes is now also best-in-class. We were now back in that market again, right, versus in the past that 7-nanometers, Avera was part of a different company where they were not as focused on the leading adds, right? So we were kind of back in the game where we can help on the custom side. But then as those conversations progressed, it also became clear they really valued a lot of the IP we had, and they were very happy to hear that we can make them available within a semi-custom product, right? And that was the genesis of essentially what you're seeing today. So you should think about this $400 million at this point, I would say it's more like a $400 million plus since we won additional designs and some of those programs. As we look forward, we can also potentially upsize Think of these as a lot of offload accelerators centered around computer networking. They are with multiple customers, and there are multiple products. A number of them are replacing existing solutions but with a lot more efficiency. So they're very important from a TCO perspective for our customers, right? So -- and I think that's what gives us the confidence that these are things which would still happen, kind of somewhat less dependent on what happens from a macro perspective. And there are some additional programs kicking on sometime next year, which will take us to that $800 million plus a year after that. And obviously, there's more activity in place, right? I mean this was the initial kind of milestone we had established. This is the first time we had talked about this new line of business for us, and this was soon after we'd acquired Avera. So it was important for us to kind of give this milestone. I mean the reality is these are going to become plus, plus, plus as time goes on, but very high confidence. In fact, a number of products for that initial $400 million ramp, we've already taped those out. Those are basically qualified, and they're going to production basically this year. So this is something which is going to start to happen fairly, fairly soon.

If you get the microphone. Yes, exactly. Excuse me. Can we get the microphone, please? Just hold on a second. She's on the way right there, please. Thank you. Just real quick on the data center on these offload kind of compute ASICs. At the end of the day, what is the advantage that you bring versus kind of like some of the more merchant silicon DPUs like either an AMD with their pensando or an NVIDIA with like what they got from Mellanox?

Yes. Good question. And by the way, I mean, we are -- sell that are also in the merchant market, right? We've been selling DPUs for a number of years. And some of them sell to the same customers, by the way, right? So I think it comes down to, again, if you've got a large enough workload where you can get an optimized solution, you can basically do things which you just simply couldn't with a standard product, right? Because a merchant product, we build as an example, it needs to fit multiple -- it's a jack of all trades, right? So it needs to fit multiple applications, which works with a number of applications and a number of customers. But I think when you look at cloud scale in particular, the volumes they drive, right, and the efficiency they can get from an optimized solution, and they can design. Remember it's not just the DPU or the chip itself. If you know you can optimize it, you can optimize the rest of the system around it. So it's not just the chip itself. It's how you actually build the whole box around it. And that gives them a significant advantage in terms of compute capability, the cost per bit, and that's what's driving this, right? And we've seen this for a number of years, right? I mean we are only now participating in this market. But if you look back, a number of the cloud customers have already done some of these solutions internally when they obviously had merchant solutions available for -- I think AI is probably the first example which comes to mind where you've got a few very large customers, which have already developed internal solution, even though there are merchant solutions available. I think it's just about optimization, given the scale and some of the workloads they're driving. Sure. Great question.

How important is 5-nanometer for the cloud optimized? I know it's important for the longer-term growth of the company. But the $400 million this year or next year and the year after, is that highly dependent upon 5-nanometer parts? Or is that kind of just ancillary? Important, but not really the driver of that growth?

5-nanometer is fairly important. I mean this was when we went from essentially a fast follower strategy. And this is not just cloud, but it's across the company to really being on the cutting edge. And that's very important because I think in Moore's Law, while I think we all agreed, the cost portion has certainly -- it's not happening anymore where the cost comes down, but there is no alternative in terms of power efficiency. You have to make the transistor smaller to switch it faster, so it's basic physics, right? And it takes less power. You just can't get around that. So for some of these chips -- and by the way, some of these chips tend to be fairly large in nature. That is still the only way you can actually get more efficiency. Yes, we also do other smarter things and how you design the chip, but you can't get away from going down the process node. So that's actually been very, very critical. A decent amount of the $400 million and $800 million is very much based on FI's.

So I think in the last question on the cloud side, you guys have talked about, I think the cloud plus 5G plus automotive having like a 20% CAGR, if I remember right. How does the cloud side fit in that hierarchy? I assume all 3 of them aren't right at 20%. Would it be faster or slower than that in just percentage terms? I know dollar-wise, it's bigger. So it's not -- percentage might not be the fairest way to think about it.

Yes. So I think at an Analyst Day, which was about a year back roughly at this point in time, I think we've shown, hey, we expect our cloud 5G and auto collectively, the multiyear CAGR to be around 20%. I think within that, the stack would be auto well above that number. I think cloud will be roughly around that number. And I would expect 5G to be perhaps in the long run be a little bit less than that number, right? I think that's probably a simple way to think about it.

Why don't we pivot over to the 5G side of things? It's amazing to me how little that comes up anymore after how that, over the last few years, was basically the only question you got, and everybody was trying to do units times ASPs times base stations and content and market share and all of that dynamic. So talk about where we are right now in that rollout. You just -- you said you hit what about $600 million annualized run rate last quarter. I think the -- if I remember right, when you have the Samsung dynamic and then Nokia eventually kicking in with the 5-nanometer side of things, that should be double where you are now, run rate-wise over the next couple of years. So just walk us through what the growth trajectory looks like over the next few years for that subsegment in your company.

Yes. I think I'll take the fact that I don't get that many questions as a sign of success, where we've gone from essentially in 4G, a single customer at Cavium. I think where the annual revenue was probably somewhere in the $100 million range, right, on average at its peak. Where today, we've got 4 large customers, right, spread across the world. We're driving a 6x kind of revenue growth rate. And we're still at the point where there's very large parts of the world which still haven't done 5G in any meaningful way, right? So I think what I would expect is first, in terms of the end market itself, the U.S., obviously, this has been a very big year. Finally, the U.S. launched in a pretty big way. But remember, no large geography can never convert fully within 1 year. So while I know the CapEx you've heard from the U.S. folks is going to go down next year. What matters to us right now since we didn't have much in 4G, what matters to us is the 4G to 5G conversion. And that's still a net positive for us even in the U.S. as we go through the next few years. That's going to be an important driver. India, finally, auctions done, very aggressively driving towards getting that up and running data this year. I would say, more meaningful would be next year. So that's still in front of us. And then Europe is, I would say, mostly in front of us, right? So I think we still have some very big regions which are still going to be converting to 5G over the next few years, right? So that's kind of a natural tailwind for our business. And on top of that, as you mentioned, we've got a second processor platform upcoming at a very key customer. So that's a significant content upgrade. And then we also have some design wins in vRAN, which is also starting to become relatively important to the industry as we go into next year. It led some of the cloud players participant in the market as well as there's this pretty strong approach for a new opportunity for 5G, which is as you know private networks right, which are probably a lot more effective than it is using WiFi. So I think there's multiple content growth in front of us, along with a general tailwind from the market still converting to 5G.

So is my math in the ZIP code that after another couple of years, once you have the second customer, maybe the vRAN side, the content per both customers potentially going up and then the geographic ramp that we should be talking about something with potentially double of where you are?

I would say we still have a lot of growth in front of us. I think that's a fair comment. While we won't handicap it more than that. But certainly, we have a lot of growth in front of us over the next few years in 5G.

For the last few years, bigger revenue growth driver areas is the smallest one today, which is the automotive side of things. That segment as a whole, I think, was kind of 5% of revenues last year, inclusive of industrial, but that's changing very quickly under the covers. Talk a little bit about what the initial products are that you're using in addressing the automotive market? And how those products are evolving going forward? Because I think the Ethernet place to start is great, but that's not the end game.

Yes. So we got into this market really on Ethernet, and this was one of the new initiatives, which Matt had pioneered after he joined Marvell, recognizing it's a great opportunity in the long run. We have the underlying technology. We just need to make some pretty strong investments to get that up and running, and that's what we did. And what you're really seeing in this market is it's an existing market. I mean there's a few billion dollars spent every year on connectivity in cars. It just happens to be on mostly older technologies kind of point to point. And as the number of sensors have expanded, right, the bandwidth is going up, the need for security is rising, I think Ethernet is kind of becoming one of those de facto standards going forward. I think our initial success, as you can imagine, was more in kind of pure EVs and hybrids, but kind of the nice part is our new design wins are now much broader in nature where they're really spreading to what I would consider the mass market. Whether we like it or not, internal combustion engines are going to be around for a while. That's going to be the meat of the market for the next few years, and that's why we are now winning designs. So our revenue, which was very small a couple of years back, I think we were at the $100 million run rate last year, we're getting towards the $200 million annualized run rate right now, and I expect that kind of cadence to continue. So I think Ethernet alone, we see the total market within the area we serve, which is primarily what I would call basic connectivity, meaning it's really from sensors to your central processing unit is Ethernet. For now within our $1 billion kind of SAM estimate, we're not counting higher speed links like some of the one-way bandwidth on a video link, right? That's like an LVDS market. Having said that, we'll see whether that becomes part of the SAM in the future. But for now, the $1 billion is clearly what's in our target, and we believe it will be at least half of the market, right, at least 50%. Certainly our design wins put us at least that much, if not better. So $200 million run rate today, I think you can expect a pretty chunky growth over the next kind of 4 to 5 years. And then the second opportunity is somewhat similar to the discussion we had earlier on kind of custom compute within cloud and the question, why do they want custom compute. You can apply the exact same logic within the automotive market, which is, look, the differentiation in cars in the long run is all going to be around your different levels of ADAS and how will you implement it, right? And then the question as an auto OEM is, do you want to outsource all of that along with all the accompanying software and firmware stack, so you want to have a piece of how that gets designed, right, and how it gets implemented. And the challenge, of course, is -- all of you would say, well, but a lot of them don't have that capability, which we agree with, which is where I think partnering up with someone -- it doesn't have to be just us, but certainly, we have the underlying flexible platform security, compute and a proven automotive architecture. So I think that could be one of the longer run opportunity. We certainly discussed one success, which is going to take some time to come to the market. But this is where we are in active conversation with multiple customers, right? Saying hey if you guys think long run. I think in the short run, obviously, they're going to have to implement it using more merchant solutions, but in the long run, if you want to control your destiny. I think this is where you're going to see a lot of interest, right? Internal investment and then partnership with semiconductor companies.

And I would think the Cavium asset at the OCTEON side with low power processing kind of offload processing, edge processing, all of those things align very well to the automotive market.

Very much so it's that. And I think now we've proven the ability to actually optimize it for multiple end market. That's important for customers too because now they've seen us take this OCTEON platform within it's original instantiation was actually really in the enterprise market. This was what powered firewalls and switches and VPNs appliances back in time. That then evolved essentially into what's now become a powerhouse in the base station market, right? And it's also now in the cloud market. So you've actually seen this. And finally, automotive would be the next -- kind of the next front here. So I think the flexibility of our platform, the ability to add and remove different accelerators for different end markets is quite frankly, I think, a unique differentiation. And we do expect that could be one of the big things why customers want to work with us.

So if we wrap all these things up, I think you guys have talked about kind of a 15% to 20% CAGR over time. Some things will be faster than that, some things will be slower. On the last call, you seem to somewhat endorse the fact that you're comfortable being within that range off of the current higher base where you're actually ahead of plan size-wise as a company. Was that meant to be a comment that was for a single year? Like, yes, next year, maybe we're at the low end of the 15% to 20%, but 15% seems right? Or was that, Matt, I think it was that made it that was just talking conceptually over the next few years off the higher base, he still thinks on average, it will be within that range.

Yes, I think it is a longer-term comment. That doesn't mean any given year couldn't be above that number or below the number. I think the context is we established this target last year, right, just about a year back. And if you look at the midpoint of the target, we should have been more towards the $5 billion run rate. We're now at $6 billion, right? So the question I was getting from investors rightly so is, hey, you're ahead of plan. Does that mean your long-term CAGR now is a little bit more difficult to get then or is the same, I don't think anything has changed. If we conceptually think about our 15% to 20%, that target essentially assumes that our end markets grow at about 10% to 11%. I think we said 11% in our Analyst Day. And then the extra 4% to 9% comes really from our own product cycles. When we look at where we are from an end market perspective, we still feel very comfortable that the end markets we have are growing at that rate, over the next, call it, few years. And then our product cycle drivers are actually probably more in the near term, maybe on the higher end of that range so I think we've talked about a number of them earlier today. So I think collectively, even off these higher basis of revenue, as you look out for our next few years, we still feel very comfortable that the long-term algorithm for us remains in the 15% to 20% range.

Got it. If we switch to the margin side in the last few minutes that we have here, the gross margin side has been remarkably consistent for you guys, considering the volatility around supply, subsegments mix, all those sorts of things. Even within pricing, you've just passed along pricing, you haven't increased the gross margin. There's nothing to be ashamed about being in the mid-60s. That's great. One of the puts and takes on the gross margin you grow within your 15% to 20% range if the price increases, cost increases from inflation foundry, all of those things normalize, will you give back to the customers? Or are there things that will make the gross margin go up or down versus that midpoint?

I think the biggest reason our gross margin is where it is, is about the capabilities we can supply to customers, right? I think, yes, we had our gross margin move up a little bit as we finished our transformation. And our goal all along was 64 to 66 and we're kind of stacked in the middle. So i think the plan has played out exactly as we expect. Yes, there were some things along the way like some of the input cost increases, but we work with our customers and we've kind of neutralized that. We didn't take advantage of them either. So we're kind of in the middle of the range. And as we look forward, we think that's still the right mix for us, right? I think the vast majority of that gross margin really is your customers value for your products, right? And I don't expect that to change. If anything, I think we're actually adding more value going forward. So I think our focus in the long run from a target model really is driving operating margin dollar growth, and we think the best way to do that is really we need solid gross margins, so we can continue to invest heavily in the business, which is what we are doing. And at the same time, right, keep revenue growth on the high side, go into new markets, which we are driving organically, CXL, AEC a number of these markets. Automotive compute in the long run, as you mentioned. I think that's what gives us the opportunity to go faster, keep our gross margins kind of where it is right now. We have been driving operating margin leverage, and I think that's going to remain the focus going forward.

Is there a wide divergence between the segments in your gross margin, the reported segments? I know there's many moving parts under the covers. But is there a hierarchy that investors should appreciate?

Yes, I think if you go back to our Analyst Day, I think that's still the right framework to use, which is consumer as well as our carrier business tends to be below our corporate average. Enterprise networking is at. Cloud, really, I should say, data center is a bit above, if you remember the bubble. And then auto industrial is the one which is well above right. So I think if you look at it, that's a pretty nice mix to have. And that's what gives us confidence that I think we'll be somewhere in that, call it, 64% to 66% target.

The last question is on the operating margin side. You guys have done a great job on that. You're pretty much approaching the range that you guys were targeting in kind of the 38% plus already. If revenue growth slows from kind of high 30s this year to, say, mid-teens, just to take the Street number for next year, how do you guys adjust OpEx?

Yes. Quite frankly, I mean, right now, our goal really is that we've got a number of new product cycles coming up. We have a new design win, which we have yet to go into production. So our view is, I think our plan right now is OpEx is going to go into next year, but grow at a rate which is still below our revenue growth, right? So we still drive leverage, but at the same time, I mean, recognize we've got very secular markets which we discussed. We've got a number of our own product cycles. These are sole-source designs. They are fundamental to our customers' ability to execute their plans. So I think our focus right now is -- and obviously, we also have to recognize we're in a high inflationary environment. So our compensation will adjust appropriately, right, keeping all the employees we need to keep. So for now, the focus is drive leverage but we're still looking at growth next year even from a headcount perspective.

Got you. Last quick one in the last few seconds we have left, you guys repurchased -- you have a bunch of cash in your balance sheet, you're below your leverage target. You repurchased shares early on in your fiscal 2Q, but then you stopped since that point. Talk a little bit about the reason for the stop and the willingness to kind of go back to your long-term capital return policy.

Yes. We actually started earlier than we planned back in Q1, right? I mean if you remember, our goal was first to get to our leverage target and then start repurchases. But given where we saw our stock and we felt it was fairly undervalued, we started a little bit earlier, quite frankly. We did a little bit a bit more at the beginning of Q2. I think your long-run assumption is absolutely valid. I think that's going to be the focus is returning really cash through buybacks. As we discussed earlier, we have most of what we need at this point in time. It's mostly organic investments. That's still a focus. I think the other thing to keep in mind is we've been running at this $500 million to $600 million of cash when we were a company about half the size. I think we've basically more than doubled our revenue, and I think we also need to add a little bit more cash. We're not talking a lot more. But I think in the very near term, I think we'll add a little bit more on our balance sheet. We've also been focused on investing in really our working capital on the inventory side very strategically. Supply still is tight, right, especially in some of our new growth areas where we've taken the decision, we're going to add some capacity in front of these big product ramps. So we don't get into some of the situations we've had last year. So I think that's been a bigger focus is to ensure we can drive revenue growth. But as we move forward, I think your assumption is very fair, which is buybacks is going to be a focus for us.

Got it. Well, Ashish, we're out of time. Thank you so much for joining us here in Las Vegas and all your insights, and congrats on the transformation.

Thanks, Ross, and thanks for having us.