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

This is Ross Seymore again. Our next company on the virtual stage, we're very pleased to have Marvell here with us. Both Jean Hu, the CFO; and Ashish Saran, the Head of Investor Relations. So thank you to both of you for attending today.

So before we get into some Marvell-specific -- well, let me put it this way. Let me start with one question that is relatively unique to Marvell. I just want to say, first and foremost, thank you for switching to an end market split in your revenues. What led to that change? And were there any surprises as you put it together as far as things growing faster or slower than you had thought?

Yes. So I'll start. Ashish can add. Firstly, thank you for hosting us, also appreciate the question. I think the way if you think about the Marvell, we have been transforming the company how we think about building Marvell is really started with the end market. So we chose data infrastructure market because that's the most attractive end market opportunities and Matt and our team, how they do the resource allocation, capital allocation really think about where we can see the highest growth opportunities across our end markets and also where we have the competitive advantage and the portfolio. We build our portfolio, both organically through investment but also acquisitions to really focus on data center, wireless, especially 5G then enterprise networking automotive as well as other opportunities. So the earnings we just announced, you can see the tremendous momentum we have built both -- Marvell organic business grew 17% in Q2, and we're guiding 20% and the Inphi business, we just closed the transaction also growing significantly. So we think it's the right time to really provide the transparency and the business drivers for our top line revenue growth. I think Ashish can add. But if you look at currently data center accounted for 40% of our revenue, is the largest business for -- across all our product lines and the cloud data center is more than 50% of that business, which is going to drive continued growth. And of course, the carrier market is a very significant portion of our portfolio too. Then enterprise networking is -- actually in Q2 grew over 40% year-over-year by expanding our market presence in that market. Automotive, we guided Q3 like a run rate of almost $100 million. So it's exciting. That's why we are providing that end market direction. So you can see where we are driving our business and where we're investing.

Yes, Ross, I don't think there were any surprises, obviously, for us, but I would say just talking to investors after we revealed the information and followed up post call, I think the surprise, I heard back was like, "Hey, we knew you guys kept saying your enterprise business is growing, but quite frankly, most people did completely know how much and how big that was." And I think now you can see that we grew that business at a 23% revenue CAGR in the last 2 years. And of course, as you probably know, most of that is organic, right? This is really the core Marvell switch and PHY business. So I think this was an ultimate proof point. But what we've actually been saying, which is our enterprise business has been growing because of our refreshed products, our position, our ability to service our customers exactly what they want and now you guys can see it in a much more open fashion. And of course, what's very exciting for investors is I think everyone recognizes we're at the very, very beginning, right, of probably a multiyear refresh cycle. We're just at the very beginning. We're starting to see some of those benefits now. But there's a lot more in front of us, right, which will be on top of our strong share position in this market.

So I want to pause on the end markets for a second and go back to 1 or 2 macro questions quickly. The supply constraints. You guys are growing very strongly, but you also deliver results that are almost exactly in line with your guidance. Some people love that consistency and predictability. Other people want to see some more upside surprises at times. Generally speaking, when we put supply into the equation, how are you viewing current either shortages or maybe a better word for you guys would be constraints?

Yes. We continue to see supply constraints for our business. We delivered a strong performance. The way we think about how we plan our business, how we guide is, we typically look at the supply situation. When we guide, we know we have the supply to support our top line revenue growth. So that's typically how we do it. And the team is really an execution machine. I think if you think about what we are facing is we are delivering strong revenue growth performance. But overall, the demand is growing faster than the supply we can support of our customers. So we do see demand continue to outpace the supply and the gap actually is widening for sure. That's really what we're seeing right now, is we are managing it, and we are setting up the long-term capacity agreement with the suppliers to ensure the supply for the future next year. So we do feel comfortable we have the supply to support the second half of revenue growth and the next year's revenue growth. But we definitely -- the demand continue to be really strong for our products, we definitely still have a gap.

Is locking in that incremental supply or significantly larger supply in the future, just willingness to give your foundry partner a duration in your orders? Is it price? Or is the commitment and partnership you have at the 5-nanometer node and beyond also helping you get supply? Because everybody wants supply right now. It's just difficult to discern who's exactly going to get it, and more importantly, why?

Yes, I can kick it off, and Jean, you can certainly add. I think the biggest reason for us is, one, certainly is the partnership, right? We have with some of our large suppliers, but it's also when you consider the nature of our business, right? They all understand where in these long secular growth cycles. We are growing very quickly, not just because the market is right now expanding rapidly. We were going to grow at this rate or close to it, irrespective of what was happening right from a short-term demand environment, right? Because we are tied into these very fundamental cycles, whether it's the conversion to 5G, it's the ongoing growth in cloud data centers. And as those suppliers look at what they think they'll do with us over the next 3 to 5 years, not just the next 1 year, they strongly believe like we do that these are very sustainable cycles, right? So I think it's more to do with the nature of our business. And yes, we are certainly -- if I had to say one good thing coming out of this constrained environment is, there's certainly a lot more willingness in the supply chain, both from our customers to us as well as us to our suppliers on sharing a long-term road map, right? Because I think people recognize we're mostly sole-sourced at our customers. It makes sense for them to tell us what they plan for the next 2 to 3 to 4 years versus only the typical 2- to 3-quarter rolling forecast. And we are extending the same thing to our suppliers. And that's working out well for us. I think the issue more is -- supply is expanding by the way. I think there's this view of supply is a problem, it is, but supply grows every single quarter. Good news, I guess, I'd argue from my perspective is that demand is growing even faster, right? And that's creating an issue. But at some point, those lines will start to converge. And we continue to grow quite nicely to that whole period. So I think we feel like yes, it's a little painful today. But over time, I think the situation will improve.

The last question I'll have before going back to the end markets is on your product road map and that partnership with TSM at the leading edge, the 5-nanometer parts. Talk a little bit about when those parts come to the markets, which end markets they address, and what sort of margin structure they will deliver relative to your current levels?

Yes. Yes. We have discussed our 5-nanometer design wins in the past -- during the last probably -- since the last Analyst Day, right. It's -- the initial 5-nanometer part actually is related to our wireless 5G business, so we talk about it whether its take out. And then later on, we have additional 5-nanometer design wins we announced during the past in the carrier market. Then we also announced the last Investor Day in the networking market for the 5-nanometer that even storage market, right? So we have a really very broad 5-nanometer design wins and the product where our team is executing. I think from a revenue perspective, Ross, as you know, our product cycle tend to be really long. So the first 5-nanometer revenue, we probably will see will be later next year. But the real momentum, the revenue ramp, actually, it will be in calendar year 2023 and then beyond, right? So we have multiple product lines across the wireless data center, even networking and the storage. We are actually leading all the competitors driving the 5-nanometer designs. From a margin perspective, it's actually -- of course, 5-nanometer is more expensive, but the way we think about the gross margin, actually, it's really related to customer-specific and market-specific, right? It's -- also it's related to -- if it's more in the part or if it's a semi customer or if it's ASIC. It tends to be the case if it's merchant, it does not matter which nanometer which not technology, you have a higher gross margin. And if it's ASICs business model, right, you tend to have a lower gross margin, but your operating margin is much higher. So we actually -- we see the momentum across all our end market, about the 5-nanometer products ramp.

So if we dive into the specific end markets and why don't we start with your biggest one, like you said, data center's 40% are now probably even a little more than that as a percentage of sales. If we unpack that into its 2 main buckets, the cloud side being about half or maybe a little more and then the on-prem side, for lack of a better term. You mentioned on your last call that the Marvell, classic Marvell cloud business was growing even faster than the Inphi business, which I think was growing better part of 30%, that's really impressive. What's driving that growth in the classic Marvell cloud subsegment of data center?

Yes. I think the biggest reason for that, Ross, is, I think we've talked about our DPUs or also known as SmartNICs, probably an easier term for investors to understand. That business has been growing quite nicely for us, and it continues to grow very strongly, right? So the trend for more and more offload continues quite strongly at our customers. And that's a big reason why our year-on-year growth rate actually accelerating within kind of, you call it, the Marvell organic business. We also have some custom do-it-yourself SSD controller wins where we directly supply a cloud OEM, and they buy NAND and they build their own drive tailored to their specifications. That is also trying to ramp basically in Q3. So that's another kind of organic business. And then Avera, when we acquired them, they actually came with some design wins in the cloud space. Some of those are now starting to kick into revenue, right? So we've actually got a lot of underlying growth drivers. And then the final one I mentioned, which has been growing continuously and continue to do that is the nearline business, right? This is SSD. HDD controllers as well as preamplifiers. I mean that's been a sustained growth driver, right, as capacity continues to grow. But I'd say, for this specific quarter, to your question, I think it's really our SmartNIC business, it's some of the ASIC ramps, it's the custom SSD controller. It's a combination of all those things.

So all those things can be lumpy in their own right, but I would describe them all as being very company-specific product cycles more so than just you benefiting from a rising tide of cloud spending. How do we think about that half of your data center business going forward? Because we all know for better or worse, the cloud side can -- as much as it gives, it can take away at different points in time. And with 40% of your business being exposed to it, there could be too much of a good thing at some point in time. But these product cycles could offset it. So how do you balance what's company-specific versus the end market just being strong and the sustainability of those product cycles?

Yes. I think the good news is I think we have both. And I will let Jean articulate it in more detail, but the answer is we've got both the lifting tide. Certainly, if you look at some of our Inphi business, right, where it is just growth in total usage of optical interconnects. But on top of that, we have a number of our own unique things, and Jean, you can also mention some of the new design wins we talked about. So I'll let you finish that talk.

Yes. I think Ashish is right. If you look at our unique product cycle. So we actually have a very broad set of products addressing cloud data center market, right? Ashish mentioned on the DPU side, not only offloader for just SmartNIC and the security offload, right? And then do-it-yourself SSD storage, which has just started ramping. On the Inphi side, right, you can see the 200-gig, 400-gig adoption is ongoing and not only the inside of the data center, the optical connectivity but also we said in Q3, we are actually going to see 400 cloud adoption and the ramp. So you do see all those product cycles with all the multiple product lines continue to ramp up because of the technology adoption because it's our new design wins. And then the ASIC, we talked about during the last earnings call, we have an additional networking ASIC win, which we think is going to come into the revenue ramp sometime later next year. So secular trends actually are the most important drivers for our business in cloud data center. So we do see cloud data center as part of the data center business will continue to expand rapidly. In the longer term, of course, the on-premise one tends to be very steady. So we can see in the cloud data center, majority of the business actually is going to come from cloud data center over the overall data center business.

Yes. So it's something, Ross, I think you're right. I think we have probably more dependence as our own design wins start to kick in, right? While we certainly benefit from the underlying tides rising. I think we actually have a significant amount of revenue in the multiple hundred millions of range coming essentially from our own product cycles over the next several years.

So the on-prem side of things, is it just a resource allocation, a relative resource allocation decision to not focus on that and focus on the cloud? Because I would think that, that on-prem side would have very similar attributes to some of the enterprise segment that you have, and you've been able to grow that very strongly. So why is there not a similar opportunity on the on-prem side? Is it just a substitution effect where it's going into cloud and taking away from on-prem?

Basically, yes. I mean that's what it really goes down to. It's a nice market, by the way, it's got good margins, right? And it's a similar products. Some of them are similar products which sell on both sides of the market. So it's one investment and you're leveraging it. It's just that the reality is if you look at the relative growth rates in cloud are just significantly higher given the use case continues to expand, right? Part of it is, to your point, it's a workload shift from enterprise to cloud, but I think there's also a much bigger piece, which is cloud just as fundamentally things which just haven't been done before, right, look at the use of AI and a number of consumer applications but simply never apply to the kind of on-premise data center market, right? These are new use cases. So I think that's why cloud has a higher long-term secular growth rate. It's not just, hey, I'm relying on growing just by sucking in workloads. I think in fact, a lot of that is kind of done. I think it's more the fact that you're just doing new business cases, new things which just never were done before.

Why don't we switch over to the carrier side of things? I believe on the last call, you talked about that roughly 20% of the company being evenly split between wireless and wireline as well. It's a similar sort of story and that half of it is going to grow really fast and the other half be relatively stable. For the 10% of total company that's wireless infrastructure, talk a little bit about the growth rates there. Any changes that you've seen? Second half of this year seems to be doing very well non-China or ex China and China growing would just be kind of icing on that cake. Any sort of updates on the end demand by geos and/or the product cycle timing when you have some of the new products to take share with some of the customers that you've announced already?

So yes, I can kick that off. And it's -- that's a market -- first, yes, it's 50-50 today, although I would say, given the relative growth rate differential wireless becomes a bigger part of carrier pretty much starting literally in this quarter, Q3 itself, right? And then that gap expands over time. If the market has done what we think it will do, which is not a big surprise, right? This is an infrastructure market where these are multi-quarter rollouts where people have to physically install base stations. So I think it's been fairly predictable. We've actually grown for multiple straight quarters. China is a relatively small part of our total exposure. So we kind of bottomed that out in Q2. We see the U.S. in particular, kicking on as we expected, and that's what's driving the Q3 and Q4 for us. And then we see that U.S. -- that's going to be a multiyear rollout over the next couple of years. Europe starts to kick in sometime next year. India probably a little bit after that. So I think this has been a business where we expected it to start to kick into a higher gear back half, and we're certainly seeing that. So I wouldn't say there's hardly any surprises. We've done a good job of securing capacity, given we call these signals well in advance. So I think we've done a good job. We've got the parts. We're happy to support our customers. We see very consistent long-term growth in this business. To your last question, yes, we also have some -- in addition to the market going from 4G to 5G, we also have an additional care similar to our cloud story where we've got some unique design wins where we're actually gaining socket share essentially. And that is a 5-nanometer part. That's the one which Jean mentioned that we've already done the development work. So you should expect, as we get into late '22 as we get into the calendar '23 time frame, that's kind of a next kicker on top, and that's playing out as we expected.

The other half of that business that's not wireless, obviously, is a smaller part, but I don't get the sense that it's as hobbled as potentially the on-prem side of data center would be. There's not the substitution effect. So same question I asked about the on-prem part of data center. Why doesn't the wired portion of your comms business act like your enterprise business and have the opportunity for product cycles?

Yes, Ross, you're absolutely right. The wired business actually is growing and continue to grow. If you think about the wired business, first, we actually leverage our investment, right, between wireless and the wired with the OCTEON processors. Secondly, the Inphi business, the Coherent DSP business, actually, if you think about the carrier market with 5G, you do see this demand actually for the first time, is getting much stronger going forward. So we actually think the wireline business compared to wireless, the growth is slower, right? But actually, you're going to see very nice growth from the wireline business. It's actually a combination between wireless and wireline. The whole carrier business is going to grow very significantly in the next several years.

Yes, we have a very strong Coherent franchise side we inherited through the Inphi acquisition. So that business is doing very well. There's also obviously ASIC opportunities in that particular market to your point around, hey, potential share gain, we certainly see that. It's -- the cadence of that market tends to be -- that's more as steady as she goes with some growth over time versus wireless, we've got this double kicker of going 4G to 5G and massive content gain. So that's why that kind of gets to spotlight. But to your point, the wired business is a great business because that thing just grows on a very steady basis. And then every once in a while, we've got this opportunity to gain some share.

So the enterprise side, and we'll be wrapping up the end market segments here in a bit. I don't want to shortchange automotive, Matt would get mad at me, but it is still pretty small as of today. The enterprise side is growing much faster than most people believe, as you said earlier, Ashish. You're doing that at a time when aggregate enterprise spending has been arguably down year-over-year for most of your competitors and most of your customers. To me, that highlights the strength of the product cycle you're in. Do you believe that when enterprise spending starts to really come back, which it seems to be showing signs of that now that, that actually further accelerates Marvell's growth in this segment? Or is it just still going to be all about your product cycles?

I think it's actually -- I mean, let's look at our most recent results, right? So if you look back the last couple of years, we showed you our enterprise networking business growing like 20% plus year-over-year, right, for x number -- 2-plus years actually. And then in Q2, we did 40% year-on-year growth in enterprise networking. So I would argue that's a combination of exactly your comment, which is a, you've got underlying growth because of your own product cycle, which is continuing. And now we're getting a kicker on top of that, right, because of the recovery in enterprise. Even in Q3, which we are guiding essentially for enterprise to grow 30% year-on-year, right, that is in excess of what we were delivering on a product cycle basis. So I think we have both cycles continuing for a while. Now once you get 2, 3 years down the line, where enterprise networking as a market, it's gone through a bigger company, now it's growing more steady basis then at that point I think our product cycle start to become more relevant, right? But right now, I think we've got both kickers for the next several years, I would say.

So why don't we pivot over a little bit to margins and cash flow and those sorts of things. If I aggregate it up, everything you've just told me about the end markets. Within data center, you're going to be more cloud and less on-prem. Within comms, you are going to be more wireless and less wired. Those seem to be things that would be headwinds for gross margins, yet you guys have delivered steady gross margins despite those trends being effective over the last year or so. So one, am I wrong that in the mix to gross margins that the headwind that those might create? And if I'm correct, what are you doing to offset it?

I think, Ross, maybe I'll give you a high-level perspective about our gross margin by end market. And so how we think about the gross margin going forward. So first, when you think about data center both on-premise and the cloud, both actually have a really good gross margin, right? Because on-premise, of course, the server connectivity we feel pretty good about the gross margin there. But even the cloud data center, we have an Inphi business, which is accretive to our gross margin overall. And also our business the DPU, SSD DIY products, even the HCD nearline in cloud data center, those are very good gross margin business. So they actually, in general, above corporate average. I think we never report by end market. So we have not talked about it. That's the first thing is we do see -- even the revenue growth is high, but the gross margin benefit to the whole company will continue from a data center business. And then on the carrier side, I think the key thing we discussed in the past is because our wireless business has small ASIC business from Avera and some of the semi customer business, which we tend to get R&D funding and R&D co-investment from customers. So it's a slightly different business model. So we do think the wireless, the 5G business will continue to have below corporate average gross margin. But the wireline side actually is very healthy gross margin, too. So when you look at the mix change overall for Marvell going forward, we feel pretty good about maintaining our strong gross margin going forward, just because the IP value we add, the portfolio with Inphi, we actually have a more comprehensive portfolio across both data center and wireless and 5G and the carrier market.

That's really helpful color, Jean, and I thank you for somewhat front-running your own analyst meeting that you're going to have because I'm sure you're going to get that hierarchy question. So if data center is above average in aggregate, it sounds like carrier might be around average, wired would be above and wireless would be below. If I made it really simple and just said by your 5 segments, what's the hierarchy of gross margins without any magnitudes on them? Would you be able to provide that.

I have to prepare my Investor Day slide on that. So just wait for a few days.

Got you. And how about on the OpEx side of things. You guys have done a superb job of managing the OpEx and delivering operating leverage. How going forward do you manage the OpEx side of the equation relative to the significant revenue growth you've had? Is it to the target operating margin that you've talked about in the past, and I know you're going to update that soon enough, too? But talk about OpEx relative to revenue growth.

Yes. Given the extraordinary opportunities, we are seeing ahead of us I would say we'll continue to be disciplined to manage our OpEx. But at the same time, we have to invest, right, is when you look at the opportunities we have, if we don't invest to get our technology leadership, we'll not be able to continue to drive the top line revenue growth. So would talk about our target model, but we feel pretty good about -- if you look at the Q2 and Q3 operating margin, it's already about 30% compared to a year ago, it was at the low 20s percentage. So we do think the model will continue to leverage. You will definitely grow OpEx much slower than the top line revenue growth. That's how you see the operating model leverage.

That's really helpful, Jean. So the last topic in the last 5 minutes that I wanted to hit on was capital allocation and/or just cash usage in general. One aspect of that, that you've been very, very active on was M&A, and you even have another pending deal with Innovium, a private switch cloud, switch vendor. So just talk about your M&A strategy, have you gotten to the point that the majority of the assets that you really wanted, the technologies you needed to add to your portfolio? Are they in there? Or is adding those sorts of assets via M&A, just going to be a core strategy for Marvell for the foreseeable future.

Yes, I appreciate the question. As you can see, right, our team has executed really well on all the acquisitions. Not only we reposition the company's portfolio to address very attractive market opportunities but also every transaction, the integration, the synergy achievement has been largely flawless. So I do think Marvell has built a machine to execute on the acquisitions. That being said, we do feel comfortable about our portfolio. We actually have the compute networking storage and the security portfolio, we can address all the opportunities. We definitely right now, just after we closed the Inphi transaction, we want to make sure the free cash flow generated to pay down some of the debt we borrowed for Inphi transaction. And once we get the target leverage ratio, we can start to return cash to shareholders if there are no other acquisition opportunities. We'll be opportunistic on acquisitions. The key driver principles for our team is it has to be strategic and targeted and the market. It has to be financially making sense. So there are a lot of criterias. The team executed on M&A transactions. I do think if there's optimistic acquisition we can do to continue to broaden our portfolio and increase our future growth opportunities. We definitely will look at it. We have the capacity to do it. But frankly, we are more focused on just generating more cash flow and get to the point we can return cash to shareholders.

So if you're in more of a digestion mode now, and I think at least I was surprised at how quickly after the close of Inphi, Innovium was announced and maybe that's the definition of opportunistic that you just mentioned. If you're in a digestion mode now, what's the target on the leverage ratio you would get to before you would start returning cash to shareholders? And is the general idea on returning cash just to offset the dilution? Or would you like to actually lower the share count?

Yes. I think our current leverage ratio is about 3.4% gross debt to EBITDA. What you see is our EBITDA is expanding significantly and quickly with our top line revenue growth. And if we can start to pay down some of the debt, the leverage ratio will quickly get down to 2, which will be great, then we can return cash to shareholders through a buyback, which is our preferred capital return methodology. And so I think that definitely that's how we think about leverage ratio and the cash returns.

Got it. Well, Jean and Ashish, we are pretty much exactly on time, maybe 30 seconds early, but why don't we wrap it up there? Thank you so much for participating virtually in the conference. We look forward to your virtual conference not too far in the distant future here, next month, I believe, to hear more about that hierarchy of gross margins, if nothing else, gene. So thanks again for participating.

Okay. Thanks.

Thanks, Ross.

Okay. Bye.