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

Great. Thanks, everyone. I'm Joe Moore from the Morgan Stanley Semiconductor Team. Happy to have the management team of Marvell here with us. Matt Murphy, Willem Meintjes and Ashish Saran; CEO, CFO and IR respectively. Just quickly, you've heard me say this 75 times. But for important disclosures, please see the Morgan Stanley Research disclosure website at morganstanley.com/researchdisclosures. If you have any questions, please reach out to your Morgan Stanley sales representative. So thank you guys for being here. I really appreciate it. And I know you're hot on the back of OFC, so we can talk about some of that later.

Maybe we could just start off generally big picture. We sort of went from this environment where you were supply constrained pretty severely so. And within a couple of quarters that turned into some inventory excess that you're burning off, which tends to happen sometimes in these types of environments. Can you just kind of talk through what you're seeing now in some of the bigger picture trends that you talked about in your earnings call last week.

Sure, Joe. Yes. No, I think having been through these cycles over the last 30 years, I would say this is not unexpected. In fact, our Head of Operations who took a new role a couple of years ago in the middle of the supply crisis, Chris Koopmans, some of you have met him. I said, "look, I'm going to give you some advice right now, it's blazing hot and it's going to turn so fast, you're not even going to know what happened to you." So like we got to be ready. I think from our standpoint, we had a very, very strong period of growth, really starting in 2020 through the end of last year, outsized revenue performance, the Inphi asset performed really well. I think the biggest thing that surprised us and certainly has led to the downward revision and kind of the adjustments is really how fast the data center unwound. And if you actually look at the rest of the business, and you go back to last year, Enterprise was a really solid performer. Actually, even in this year, carrier, 5G is going to grow again, automotive is going to grow. So a lot of stuff in the portfolio is working. But clearly, the shift fairly quickly from the enterprise on-prem and cloud hyperscale data center that shift and that deceleration of growth was a lot faster than I think we or a lot of people had anticipated.

Yes, that's certainly something we've heard a lot of. Can you talk about where you are in the bigger kind of journey here that you started with the storage enterprise networking portfolio, you've obviously acquired a lot of capability that takes you into cloud and comm infrastructure in a much bigger way, you have building blocks now that you can go to customers and say, "Hey, we can do a lot of different things for you." Where are you with that? Do you have the IP that you need to achieve or bigger objectives, do you see further M&A in your future? Just maybe talk to a bigger picture strategy?

Yes. No, I think I'd go back to kind of the refrain we had back in October of '21 at our Analyst Day, which is, look, we did what we needed to do to get the portfolio that we wanted and we've got that in place, and we feel very good about that. In terms of all the key pieces you pointed out, the end market representation is much broader now, but also just the technology set, right, from storage to high-performance processing, which really came from Cavium, the custom ASIC capability that kind of goes across all the businesses now from Avera, bolstering the mixed signal type of stuff, with the Aquantia and Inphi and then having the lever on the optics transition. I mean, I think it really is a complete technology platform built on and that's the way we think about it as a company, right? Whether it's the 5-nanometer process or whether it's our packaging strategy, it's really to leverage this breadth of IP to those end markets, which really require high-performance data infrastructure silicon.

I mean, it seems like it's been very -- for a company that's done that much inorganic activity, it's been done very thoughtfully. It's been done with a view of what that portfolio looks like. And I guess one of the things that I've always been impressed with is the execution around those deals in that I think at your last Analyst Day, you had the founder of Cavium on stage, you had the founder of Inphi on stage. And we don't always see that with these. I don't take that for granted that you sort of keep a lot of the key people. What's been the secret to kind of unlocking value from these acquisitions?

Yes. I think it's 2 aspects. I think you touched on one very important one, which was our approach to people. I think that some of these companies were very, very unique assets, kind of crown jewels for what they did. And so our view was you can't keep everybody, but certainly, there's very special people that were involved in creating these companies. So to your point, Raghib is still an invaluable asset for us. He was a co-founder of Cavium. Loi Nguyen who runs all of our Optical Business. He's a founder of Inphi -- actually Puneet who was the co-founder of Innovium is on the team. He's running all of our data center architecture. I mean, you can -- we kept one of the technical co-founders for many years of Aquantia, and he went off to go do his own startup, which we sort of helped him with. So we tend to have that type of approach. At the same time, I would say, we were able to execute all of these deals with, I'd say, pretty meaningful cost and revenue synergies to make it work. So we weren't afraid to go kind of take out where we needed to where it made sense and do that aggressively. I think we overachieved on every deal we did in terms of the cost synergies we communicated, integrated ERPs really fast. But also I think the revenue synergies is actually really what's given the boost. And I think having all these pieces together has really unlocked I think the potential value of the whole portfolio. And I would just finally say we really saw that accelerate. I'd say even in the pendency of the Inphi deal closing that whole customer set, especially in the cloud really opened up. And I think designs, we were sort of crossing our fingers we might get basically just start to fall in our way because I think our plan, our thoughtful plan about how to integrate that asset and combined forces, it really resonated. And also the scale, right? At some point, customers are saying, "Okay, you guys are a $30 billion market cap or $40 billion market cap or whatever it is." And that's solid, right? And you're big enough and you're strong enough and their view is you're -- we don't have a scale problem per se. So we can trust you with our most critical stuff.

Yes. All right. That's great. And I would also, just from my perspective, it seems like the financial execution around those deals has been really good as well. It's the synergies that you talked about, but also just taking businesses that were sort of incented by revenue and moving the incentive structure a little bit more to design wins and things like that.

Yes, that was part of the pivot. I would say really on one of them, we leaned a lot harder relative to growth versus sort of how do you split the baby on the cost synergies, and that was Inphi. But at the same time that was done in a period of very high growth, paid on multiple, low interest rates and the strategic aspect of it was so compelling. And even though that was a bigger risk and a little bit of a departure relative to the high multiple paid, I think it was appropriate for the time, and it just absolutely delivered, right? I think outperformed year 1, it's outperforming since we've owned it and couldn't be happier. And now -- and look, it was accretive in the first quarter. At one point, we sort of stared at it and said, when is this going to be -- so I think all that's worked pretty well. But as you think about the environment we're in now, and again, we've got the portfolio we want, but to the extent we ever started looking at other things that would really be the traditional way we look at it. Really, what's the -- where can you drive meaningful cost synergies? Where can you really take an asset and plug it into our platform and repeat that playbook and then drive value both on the revenue and the cost synergy side. And not -- the era of stretching on multiple is sort of over, I think, in a large way for tech M&A.

All right. That's great. So maybe if we talk a little bit about some of the businesses now, starting with data center and maybe starting with optical because you did have OFC this week. Can you sort of talk about some of the new areas where you've seen product introductions? And maybe just give us a competitive outlook on optical?

Sure. Yes. And I kind of reflect that. I was at OFC yesterday. I don't know if some of you guys were down there or not, but spent a whole day there kind of from Sunday night with the customer -- I'm sorry, Monday night and then all day yesterday. So give you some reflection. But I've been attending that show for probably over 10 years, not every year, but I've gone -- and first of all, I'd say that there's always fireworks. There's always a rattlesnake lurking and there's always some press release and a lot of noise, and you have to sort of sift through it and figure out what's happening. I would say my reflections are we are just incredibly well positioned in the optical area. I think the first sort of big breakthrough was we were the only company to be able to show working silicon demo and a broad module ecosystem built on the next-generation optics platform, which is the 1.6 terabit generation, which also by the way, which has native 200 gig SerDes, IO. And that's sort of an episodic leap, right? I mean, you went from sort of 25-gig NRZ when PAM hit and you went to 50 gig per lane, you had to sort of implement these more complex DSPs. That's really where Inphi just sort of shot into prominence. And then it went to 100 gig and now 200 gig, which gets just vastly more complicated with respect to the DSP and the signal integrity issues and so forth. But it's really what's needed to enable next-generation AI, the next wave of high performance switches. And by the way, since it's native 200 per lane, we are also demoing an 800-gig solution that's like half the power or 30% less power than what's already out there just because now you can go 4 by 200 versus 8 by 100. So I'd say from a product standpoint -- and that was all very well validated. The theme of all the meetings I had, and it was with a range of people from directly with hyperscale folks to module partners ecosystem and at varying levels of management. The big ask was go faster, get your 51.2 T platform out faster. Our AEC demos, as an example, I mean, that's a real trend that's happening. And again, we're intercepting that market with our PAM technology from 200 to 800 gig, demoing all of those solutions with us and our partners. We've got every basically cable manufacturer partnered with us. So real strong demand and I think the idea that pluggables are here for the foreseeable future is definitely the refrain. Certainly, we're aware that there's always different architectural solutions that get proposed, whether it's co-packaged optics or linear drive. Some of these have been around since I started to go to OFC in 2012, and it was sort of like it's coming, it's coming. So I think we feel really good about where we sit competitively, strong customer feedback on the total solution and discussions to your point, Joe, about more broadly, especially with the cloud type of companies, how can we do more together relative to that whole portfolio with physical layer leadership, ASIC capability, switching and thinking about next-generation AI applications and things like that. So a lot of real positive -- real positive show.

Yes. I mean it seems really strong. And I know at one point for [Indiscernible] Had talked about we're going to maintain 50% share over time. And it seems like that was pretty significantly conservative based on what [indiscernible]

It was. Yes.

Can you talk about that in the context of hyperscale budgets coming down? I mean, it seems like this is going to be an area that's somewhat protected, but even NVIDIA had issues last quarter. So I mean, which is obviously in a protected area as well. But how do you see that tug of war between -- the budgets are coming down, but this is an area that's key to these new workloads that we see coming.

Yes. I think budgets coming down for sure. I mean, I think some of our customers are here. So they're probably talking about OpEx and CapEx and how they're going to go extend the life of platforms longer and get more efficiency. I think that started kind of at the end of last year, and that's continuing. But it's definitely not an if, they're going to upgrade. It's just the timing of which I think this if you think about even like one example, right? The upgrade -- the networking upgrade to 51.2, I mean that's going to drive a 4x increase in network capacity, right? And by the way, they're not going to do this asynchronously. They're going to want to make sure everything is ready from the AECs to the next, to the switch, to the optics. And so that's coming, I think -- and then I'd also say that there's other things that are happening independent of budget and AI would be one of those. I mean, there's a lot of talk about 800 gig, and it's coming in the future. I mean, we went to massive high-volume production last year on 800 gig. But that was for AI. That wasn't for switches, right? I mean this was for AI cluster applications. And that obviously has had no budget constraint because that's actually where everybody is putting their energy. And the final thing I would just say is that even some of the upgrade from NRZ to PAM still isn't done. I mean, we had 1 of our customers talking to us about, look, we were a little bit slow and there's been some delay to go from sort of 100 gig to 400 gig, but that's actually happening next year. So while there's been some delays, and certainly, there's inventory adjustment that we're going through right now, our module partners were bemoaning their inventory issues as well. But everyone sort of sees the path out of it and the higher level, where the puck is headed, especially into 2024 is just very, very strong from an optics standpoint and the whole ecosystem around it.

Yes. And then the other growth driver -- big growth driver that you have in data center will be cloud custom silicon. I know you've pushed the time frame out a little bit on the financial targets there. But can you kind of talk to those opportunities. And I think with as much specificity as you can give us kind of what types of programs are those? And I know there's custom activity in every part of your business, in optical there is, in DPU there is, in storage there is. But it feels like you're talking about here more compute offload stuff that's incremental to those opportunities. Can you talk to that?

Yes, maybe on that part. And I think as we get closer, I'm hoping we can talk in a little more specificity. But as you can understand, Joe, given the confidential nature of these and the fact that these products are basically what -- even if we're building the whole chip, it's their product. It's going to have their logo. So I think there's a lot of sensitivity. But I would say, generally, -- it's a diverse set of applications, but many of them and most of them are compute-intensive, and that really plays to our strengths from our ARM CPU capability, which we got from Cavium and have sort of advanced along the way. Relative to the timing, yes, it's pushed out by a few quarters. I think the one thing I regret is we've done a pretty good job and have a good track record, I think, of being able to size new growth, the opportunities for investors and then kind of bringing them along. We started talking about automotive a few years ago. It was -- there's no revenue. It's organic, ground up, and we said, "Hey, we think this is a $100 million business." And back we were like $2 billion and change in revenue. That's a pretty big deal. And then we got there, right? And then we kind of moved the goalpost to $200 million, and we hit that last year, and now we're saying it can go to $500 million. And so I think 5G, we did something similar. The difference on the cloud one, which in retrospect, I would not have done is we've -- for the first time, put a time bound on it, which was, hey, in this 4 quarters, this is what's going to happen; this 4 quarters and -- looking back, obviously, we had not anticipated exactly when that was going to happen correctly. And so it's moved out. But that's more of a function of what you said earlier, which is really just trying to sweat it out with the existing platforms longer that they have, and that's sort of their corporate decision. But we still view these as very compelling part of their next architecture and mission-critical to what they do. So probably not going to happen all this year. We said just take about half of it and move it out. But even then, we're we've got to really see how the year unfolds, Joe. It's been pretty dynamic, so...

I mean in your defense, every next-generation server platform has sort of seen the same thing...

Correct. Several platforms themselves have moved out. Some of those were attached to that cycle in the context of, "Hey, we want everything synchronized." There's a whole thing like a server platform just moves out 6 months or 9 months, you're probably going to get drug with it.

And then as we segue to enterprise networking, but I mean just more generally, as I think about these inventory corrections in data center networking, how much of the weakness you're seeing is you kind of proactively cleaning it up and trying to get ahead of it and reduce inventory so that you can have a better backdrop in the back of...

You're talking about enterprise?

Yes, primarily. Data center too, like to the extent -- I feel like some of this is the market being weak and some of this is Marvell kind of actively making sure that, that inventory gets cleared out.

Yes. That's generally, to your point, it's a broader business. So geographically, it's dispersed. Last year, we saw China really start to inflect negatively, right? That's sort of their own issues in that country. And again, we kind of let that inventory float down. So a lot of that goes through the channel, right? So that's an opportunity to work with the customer channel partner versus some high-volume mega socket, that's one customer, one deal, and then it's a little bit harder to manage. So yes, so very much so at the end of -- in fact, I think in our December call, right Ashish? I mean, we talked about the fact that, hey, look, for the China enterprises -- look, it was -- enterprise for us was growing like 50% a year every quarter annually for a long time, okay? And I think we were pretty transparent that, look, we think probably this calendar year, there's going to have to be some level of reset inventory correction. And so to your point, yes, we're actually trying to do that more, I think, prudently and aggressively now and get the channel in the right place, the customer is in the right place, let them schedule around what they need so that we have a much better picture of supply and demand and quite frankly, have a better setup for the second half. And this is in the context too -- look, supply has opened up, lead times are coming down. We also want to get those over time back into a more normal place. So when you go through that, my experience is, it's a little bit better to just get it flushed and get it cleaned so you have a better baseline that you can grow off of. And that's how we've thought about it. But at the same time, we're very mindful of people placed orders that are not cancelable. So okay, we've got to take it, but -- let's figure out a way to try to get that to match your demand a little bit better. And we think that that's a better -- my experience that's rather do that for the long term, make that decision now.

I mean that 50% growth profile that you've talked about that you've maintained, -- there's -- that's actually been going up for quite some time and you go back to the Cavium days and their business and enterprise networking would significantly outpace their customers' growth. But it's also coincided with a large inventory build, large customer prepayments and things like that, that happened in that space. So can you kind of give us a sense for what that growth trajectory looks like over the longer haul? And how long it takes that inventory to clear out and what the growth looks like thereafter?

Yes. I think that, to me, looks like a more normal traditional supply/demand inventory kind of rebalancing -- commentary, quite frankly, out of the North American companies has been actually surprisingly generally positive. If you look at their view of the market. They're not saying it's growing gangbusters, but they're also not signaling a major, major impact. And then I think on the China side, by the way, it was so low -- my opinion is it's going to come back for sure, especially after this party congress is over. I think there's going to be tenders and bids. I think they're going to resuscitate their economy a little bit. So we're certainly hopeful by the back half that, that comes back. And then over the longer term, we've always -- I know we look like we just totally under called it, which we did, but we've always said, look, we think enterprise is kind of a GDP as a market grower. And then we can outgrow that. Look, I think when we said it was 4% or 5%, we thought we could do a little better. And of course, we did 40 or 50, so got it, missed, okay, things went better, shoot us for that. But I do think it probably goes back to okay with that. That's okay. We'll take it because actually, we stepped up to basically a new base in terms of -- and we'll take it. It's merchant products for the most part that got leveraged across a whole bunch of companies, and we gain share, right? So -- and that's obviously the leverage you get because you're not -- the R&D leverage is quite high. So I think over the long term, it will be a grower for us, but it's not going to be at that rate. But certainly, can we outgrow the market. I think that's -- once things are normalized, I think that that's a reasonable assumption for people's models. I wouldn't say anything has changed there. We just kind of keep chunking away at it.

Yes. All right. And then one of the -- you have a couple of growth stories that are working really well this year in carrier infrastructure and automotive. Maybe starting on the carrier side. You talked about 25% sequential growth in 5G infrastructure this quarter. That surprised me a little, it's not -- I think that reflects maybe the strength of some of your 5G programs. But can you talk to the drivers of that growth?

Yes. I think a couple of things are going on. The first is, I would just say, and every seasoned investor knows this, which is the famous quote, "carrier is lumpy." It's just -- and so -- normally, people say "carriers lumpy when it's down." And they say why carrier is lumpy -- sorry -- but actually, we're saying, look, carriers is not. So it's got a head of steam on it, which is great. Part of it, too, why it looks a little different maybe, I think, than kind of some broader industry commentary is. You got to remember, we didn't have a whole lot of 4G content at Marvell, right? We really hit the 5G inflection. So when people talk about comms, there's kind of that offsetting issue, legacy equipment going down, 5G is where all the CapEx is going. So that may be a little bit of it. You've got India, which is very aggressively finally rolling out 5G networks in a very big way this year. So that's part of it. And then I think some of our content gain we've gotten over the years where our customers are introducing the new older platforms are rolling off, new ones are kicking in with our solution versus what the incumbent was. That's probably some of it, Joe. And I think it probably has a very strong, obviously for Q1 and probably continues again in Q2, and then we got to kind of see where it goes. But overall, net for the year, we do see 5G up again. And we're very pleased with the trajectory of that business. It exceeded our original $600 million target. It kind of got there last year. It's going to grow again this year. And then over time, it will moderate, but there's still -- there's still some content gains that are in front of us, and it's been a very good business for us.

Yes, Joe, I mean 1 thing to add also to your 25% is, our customer base is a lot more diverse today. And those customers are doing well, quite frankly, right? If you look at the share gains that are executing on, we're seeing those benefits, right? As we kind of ramp and then what we still see in front of us is another very important socket we've talked about, right? Which is completely in front of us, right? So I think we can continue this growth for a while.

Yes, I do think the impact -- I think there's even an accurate acronym for this now that some of the base station company is called NRVs, nonreliable vendor. So I think there's an NRV kind of rip and replace going on in Europe, certainly in other parts of the world, and that basically, that share is actually starting to move and ascribe to some of our reliable vendors, I guess, the operators would say. And so that's probably a little bit of it, too. But look, we'll take it. We'll take it in a choppy world, and we'll take some growth right now.

I'm discouraged that, that vendor is still not reliable -- another conversation. And then I do only some time for the financial questions. But automotive, can you just talk to the growth drivers there, Automotive Ethernet and what are the prospects for expanding the scope of your Automotive business over time?

Sure. I mean, I think this is a great story for us, right? As much as sometimes we can get easily characterized as well, Marvell is just a roll up, you just bought these companies. You didn't put any -- I mean, the success of our organic investments, really where we thought the portfolio had some potential, which was pretty thin, slim pickings, quite frankly, the original Marvell if you just sort of unpacked it. But where we put energy, it's done well, and Automotive was a great example. That business is going to grow again this year for us. It was about $200 million last year. The exciting part about that, Joe, is the design win activity is been through the roof in the last 2 years. And to a level where the lifetime revenue we're winning is so far in excess of the revenue we've got that it gives us a lot of encouragement for where this is going. And one of the reasons is the initial deployments of this were either in the case of traditional kind of combustion engine vehicles, it was more of a one-off link here there, that's the last few years. And then the EV guys just went all in on it because that was sort of -- if you could go ground up why would you not actually pick a highly reliable Ethernet backbone to connect everything in the car, right? So content was very different on those. EVs are doing really well. I mean, that's just been a share gainer. So that's helped us. But the game changer has been that all of the major traditional automotive companies with big combustion engine platforms are also moving aggressively to the concept of smart vehicle, connected vehicle, et cetera. And so those new platforms that we see coming -- and you see more of the hype comes from some of the ADAS companies, right? Where they're winning here and there. But we partner with everybody, Joe, we're like -- we're Switzerland on this one, right? So we don't -- we basically -- we're in all the platforms. And so when the traditional gas-powered vehicles start rolling this in, it's just a huge content gain for us. and it just opens up a larger number of vehicles. We invested to win. We -- this was a negative P&L for a while. We put a lot of energy into it. We believe, and now it's a profitable business where we can actually kind of -- it's in a great place because it's growing every year, so we can add R&D and keep funding it and -- so it's -- we've turned it from a startup to a real business. And I think over the next 10 years, networking and automotive will be just a very, very important trend. I think that Marvell will be a leader in.

Right. That's good to hear. Maybe touching on financials. I mean, one of the surprises out of the quarter was the gross margin hiccup in the April quarter. Can you talk to the drivers of that? And then how you -- you talked about that getting back to model over the next few quarters. Can you talk about what it takes to kind of get back there?

Yes. Maybe I'll let -- since we gave Willem an amazing gift this first quarter as a CFO is a wonderful gross margin guide. Willem, why don't you give some of your thoughts on that.

Yes. So clearly, the primary driver there was the mix, right? And specifically if we start in data center, clearly, through another step down, leg down in storage. And then that was a bigger step down than the general data center product, but we saw a decrease in that as well as the inventory correction sort of broadened and it was something deeper than we had expected in Q4. And then in addition to that, we saw on the carrier side, weakness in the wired business, and that's typically a higher-margin business. And then that was offset by the 25% growth that we were talking about in 5G. And then when we look at the merchant networking silicon, that was done as well. And it was really offset by ASIC business that we're starting to do well. So when we look at the sum of that and the lower revenue, certainly, we had some lower absorption as well. But as we look at the rest of the year, at some point, we do get through this inventory correction, our mix will start normalizing. And so we're fairly confident that exiting the year, we'll be close to the to the bottom end of our long-term range of [46 -- 64 to 66] , Joe.

But there's no areas where there's like-for-like pricing differences or...

Yes, really not, right? Our business is really long term in nature. We have long-term contracts and certainly nothing like that where we just change pricing or had some volume discounts or anything like that.

And then maybe just to wrap up because we have only 1 minute left, so much to talk about with you guys. The -- can you talk to your financial objectives going through a macro rough patch and how do you protect the investments that you want to have versus optimizing earnings in...

Yes. Yes, maybe 2 aspects to that just quickly. When we look back at last year, one of the things that's going to be my primary focus is free cash flow, Joe. And so when we look at last year, we were really investing in working capital in some of these supply chain agreements, right? And so that piece is also behind us. So there's an opportunity for us to really focus on getting inventory lower, improving ideas, so really improving free cash flow. And so we're in the near term, focused on repaying this June [$500 billion] bond. But once we get through that, clearly, there's an opportunity for us. And we're excited about starting buybacks again. And our capital return policy is actually to do a minimum of 50% through dividends and buybacks. And so in the second half of the year, I think if we're able to do more, we're absolutely fixable around that. And then I think the second part of that question was around capital allocation. And so we're clearly -- we have all these design wins at our customers are relying on us on, and we're really focused on executing on those first and foremost. Clearly near term it's pretty choppy. And so we're hyper-focused on controlling OpEx. But then also going through our capital allocation process and as far as the opportunities and projects that are or maybe slightly delayed to just tweak those investments and optimizing. And so that's sort of that OpEx guide that we gave towards the end of the year.

Great. Well, thank you very much. We'll wrap it up there. Thank you guys.

Thank you Joe.

Thanks Joe.