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

Good afternoon, everyone. Thanks for joining us on this day 3 of the BofA Global Technology Conference. I'm so delighted to have the management team from Marvell join us, Matt Murphy, the CEO; and Ashish Saran, the Head of Investor Relations. And what we will do is perhaps, I'll just pass it on to Matt to really give us kind of a summary, right, kind of a state of the union, right? Matt has been there for now 5 years, and the company has just gone through very remarkable transformation in that timeframe. So I think it will be very beneficial to hear from Matt what the company has gone through. But then importantly, how it is positioned in this new computing market, what the addition of Inphi means to the story. But let me pass it to Matt to introduce the company and how it's positioned. Thank you, Matt.

Sure. Yes. Thanks, Vivek, for hosting us. It's an interesting milestone to reflect on certainly, any time you have a 1-year, 5-year anniversary type things. So yes, it was -- actually, I was looking it up, it was June 20, 2016, I was announced as the incoming CEO. I left my job at Maxim where I was at for 22 years, and I just left on July 11. And it's interesting. We're -- Vivek and I were chatting earlier as we were getting ready to come on. I mean, if you look at sort of where Marvell was at that time, notwithstanding the financial concern, the sort of founder changes, activists, all the noise, the company positioning was in a really tough spot. And I think my first year of earnings calls, even for a few quarters, was really about, well, how much of your storage is HDD versus SSD? And you have this melting ice cube, and what's going to happen. And you have all these consumer businesses, what are you going to do with it? And enterprise isn't going to grow. Anyway, there's like a lot of concerns. And so, I think if you just sort of look at the journey we've been on, there's really 3 vectors to it. I think the first was, probably most importantly at the beginning, was what we decided not to do, and that was really the editing that we did early on about divesting noncore businesses, shutting down a bunch of internal development that was into the consumer market. The second was really double downing on our organic investments in storage and in networking, which I'm actually very pleased with the results of, and Vivek, we can talk about that later. But If you were to say 5 years later, you would have a Marvell storage franchise that was actually growing double digits that was heavily concentrated in data center and cloud, I think, nobody would have believed that, that was sort of possible, and we've achieved that. And then, in our organic Marvell networking business, that thing has been absolutely on fire, right, the last few years. And even in a flat sort of growth market, we've been growing that business well over double digits. So the organic investments should never be overlooked. And then of course, the headline news is always the M&A. And those were -- each one of those very important transactions for us. Cavium was the first one, which really got us the core processing assets based on the ARM architecture for high-end CPUs. We did Aquantia to fill out our Ethernet portfolio and then Avera for custom ASIC, and then finally, Inphi, which closed in our last quarter, which we got 10 days of revenue on. And so, those also have repositioned us. And so, if you sort of look at like a sell-side note from 5 years ago, it would have been around HDDs and Wi-Fi and multimedia chips. And now, it's about 5G base station infrastructure, cloud, custom ASICs, high-end 200, 400, 800 gig optical, do-it-yourself flash array storage controllers for the cloud. So it's just a really exciting portfolio we've put together. And that's why I said in my -- in the conference call recently, I really do believe this company's best days are ahead of us. And all the stuff we done over the last 5 years, which has created a lot of value, I think there's probably multiples of that to be had relative to the bigger opportunity, if I look out over the next 5 to 10 years, candidly.

Excellent. Thank you for that introduction, Matt. So what I hope we could do is perhaps just go through some of the supply side issues first and then go through the big demand drivers. So on the supply side, just widespread constraints in the industry. But we have seen differences. If you happen to have a bigger consumer footprint, smartphone 5G, than perhaps you're less restricted. If you happen to have kind of more complex products like [indiscernible], like there's a lot more restrictions. So describe to us how constrained you are from a supply perspective? What are you doing to alleviate these constraints? And then, most importantly, when do you think you will be able to get back to trend both in terms of kind of quoting your normal lead times?

Yes. Great, great question. I think you highlighted a key point, Vivek, which is the following, which is -- remember, this is a $500 billion global industry, and not all semiconductors are made alike. Now Vivek, you would understand this very clearly, but a lot of people who are more generalists in tech that are kind of learning about semis may not realize this, but you have products today in certain parts of the chip industry that aren't constrained, right? If you have your own old analog fab and you have parts that have 16 pins and you have a package that has 5 sources, you could probably ship a bunch of upsides right now. And there's 4 sources for these commodity parts. But then you go all the way to the bleeding edge where you have very complex substrates and manufacturing processes, and those are much more scarce. So there's no single answer to "when the shortage is over." I think that's maybe the most obvious statement ever, but I don't think people realize that. So when -- by the way, when it looks like CEOs aren't aligned on this, some say, "Oh, I'll be done. I'll be good in one quarter." "Oh, I'm at the peak of the cycle." Others say, it's going to last 2 more years." I don't think anybody is lying. I think, every company has its own set of dynamics. So for Marvell, I guess, here's the good news, bad news, right? I mean, we've been growing this company, Vivek, at well over double digits since a year ago, right? Even through the COVID pandemic, Marvell was growing very solid double digits, maybe call it low double digits. Our Q1 we just announced, we actually grew Marvell 17% organically year-over-year. So that growth rate has actually been accelerating. And then, what I said in my prepared remarks is that in the second half, that growth rate percentage would accelerate further. So we're getting supply, okay? And we've been actually -- I know we've gotten ding for it in the past, but we didn't have the dip and then the -- we didn't have that period where we could even build any inventory. So we're pretty pleased. We've been delivering to the Street expectations since that time. We're exceeding in terms of revenue. And then we had a nice beat actually on the EPS line and a guide above. So I think that's all good. Now what I told my team just a couple of days ago on our employee all-hands after the earnings is, I said, just because we're meeting the expectations that we guided doesn't mean I'm satisfied. And so, to your point, it's really about customer satisfaction. It's about getting your lead times back under control. And I guess, the good news is, business is so good that we've been able to grow the revenue. The bad news is, we haven't been able to make a real meaningful dent in the delinquency, in the backlog that's unsupported because that means the demand is continuing to grow faster than even we can supply it, even at these rates. And that's very different, by the way. If you kind of unpack us versus that, if you just took our product set versus some of our peers, I think you'd find we're performing extremely well. So we have a number of initiatives in place, Vivek, to address this. The first is we did make a leadership change, and part of that was really around -- in operations. And that was really around instituting a completely different sort of strategic framework for how we manage our supply chain, long-term contracts, improving our internal forecasting systems and tools and processes, so we can do multiyear capacity planning. My engagement, over the last 6 months with our supply chain, all the way down to the substrate level and the CEO level of all of our subcomponent suppliers, is really good. We're telling our story, just like I'm telling it to our investors. And we're using our balance sheet, in some cases, to make investments, to make sure we get the capacity we need. And as I said, we're seeing improvements, even in the back half of this year in terms of our supply to enable us to get there. When we get back to "normal lead times," that's unclear to me. I think, we're still and our industry peers that are doing well are still, especially when you look at the complex substrates and some of the pinch points that are out there, it's not like clear exactly when that frees up. So we're going to just keep focusing on quarter-after-quarter and also lining up the longer term. I mean, I'll just tell you, even -- and I'll pause now in just one second, but we mentioned in our prepared remarks in the earnings call that we had won a number of new pretty important designs in the data center that we're going to ramp kind of in '23, '24, '25, that would be on top of the growth we're already seeing. I'm already in direct discussions with the entire supply chain, okay, about how we're going to support those products, units, package sizes, volumes. And we're -- we've pulled all of that decision making up to literally when we do the plan of record approval for the project. I mean, just to let you know, most companies, the way they do it is the chips done, it's taped out, it's in the fab, and then the ops teams run around and say, okay, let's quote it to 4 vendors and see who comes back with the best price. It's like a complete rewrite of how we're thinking about it because we can't afford to not be prepared. And so we're prepared to make those commitments now because those projects have NRE. They have contracts behind them, and they're not like brand-new science experiment types of chips. They're either follow on to what's already being done or replacing an incumbent. So we feel pretty good about where we stand there, Vivek, and we're going to get through it, but I think the whole industry has got ways to go actually before it's out of the woods.

Got it. And Matt, one other thing you mentioned. You have made some leadership changes on the operations and the supply chain side. So talk to us about that because my impression is that as Marvell was kind of transforming over the last few years, it seems the company was making sure that there was a very tight control on the cost side, right? There was a lot of discipline that went into that. And now, you have this kind of really V-shaped recovery, right, or V-shaped just growth in multiple dimensions. So how are you ensuring that you maintain that discipline while not leaving a lot of upside, right, on the table?

Yes. Well, I think you nailed it, which was the environment 5 years ago was -- and candidly, the value creation opportunity in Marvell, as I saw it when I became CEO, was actually through cost rationalization, more efficiency, better run company, and we did that, by the way, right? Like that created -- I mean, the first self-help we announced was like $200 million, right? It was just like we just did that on our own. And then, if you remember that gross margin step up, it was sort of a sub-50% gross margin company in 2015 or in that range. And we got it right for Cavium. We had it in the in the low, almost mid-60s, right? So that -- so the good news is, Vivek, all those financial controls, measurements, metrics, that's all -- those dashboards are built, right? That's how we still monitor the business. But you're completely right, is that when you go from that kind of cut mode to growth mode, in so many aspects of running your company, you actually have to rethink and reimagine it. And I have -- we're not losing any kind of recipe on that. But I think, we definitely went through that pretty fast. It was not actually real gradual, right, coming off of trade wars and M&As and some divestitures. We kind of went through this period where, all of a sudden, we went from very low kind of almost flat growth with the ins and outs to 17% a year with Inphi going faster than that, right? So that change has been fairly quick. And so, I think, our ability to pivot on that as crisply as we needed to didn't happen to my satisfaction. We tried, but I think, with Chris at the helm -- and by the way, if he's been with us since day 1 and he ran -- he was actually Vice President of Worldwide Sales, he was Vice President of Networking business, right? He handed that over to Raghib when we bought Cavium. He ran all of my corporate M&A, PMO, all the -- like number of company initiatives, engineering program management, he's extremely operationally astute. And the advantage of actually having someone like him in that role is he knows the product, he knows the technology. He, knows the circuits, he knows the packaging, he knows -- and he knows the demand, and it's going to prove to be a very valuable asset. So yes, you've -- but we will clearly not lose the recipe on mark. And looking, even in the short term, our margins are actually hanging in there pretty solid despite some of the wrong ways on input cost. So I think, the whole team is going to manage it well. It's a good combination.

Yes. I mean -- and it would be strategic in how we do this, right, because if you think about the nature of our products, majority -- vast majority are sole-source proprietary and loan life cycles, given our infrastructure focus. So I think, we can be pretty smart about, hey, there's a few areas where you can actually afford to say, use your balance sheet, be better prepared to meet upsides from customers without necessarily getting stuff or taking on extra costs. I think, Chris is -- that's one of the key things he's doing, is strategically figuring out, hey, where are some areas that we can afford to be more flexible, I think.

Got it. Absolutely. Now one interesting thing, Matt, that you mentioned, is that you are expecting your sales growth to accelerate in the back half. And I know, when we hear it, we say, okay, impressive. But when you actually look at it in the context of the semiconductor industry, semiconductor industry sales growth will actually decelerate in the second half, right, because we are kind of past the easiest quarter of compares. So talk to us about what does this acceleration mean. If you can quantify it, that will be even helpful. And how much of this is organic because of organic product cycles? How much of this is because Inphi is now getting layered in? And from what I recall, Inphi, when you reported, right, is growing even faster, right. Low mid-20s kind of year-on-year growth rate.

Yes. I'll answer it, and then I'll have Ashish at it as well. So I think, a couple of comments. So the first is, yes, we see a path as Marvell stand-alone, okay, ex-Inphi, to improve the sales velocity in the second half. And then the reason is, well, why is that? And by the way, I don't think that -- I don't believe that's a second half phenomena. I mean, we actually have very good line of sight into calendar '22 or our fiscal '23. And again, this is all design win, product cycle driven from Marvell, right? That's the -- I'll get to Inphi in a minute. So just think of the Marvell business. The first is you have acceleration in 5G in the second half. I mean, if you think about now, we just recorded our seventh straight quarter of growth sequential in 5G despite the whole China pause now. For Q2, we didn't guide 5G up. But certainly, we said for the second half, it's going to come back pretty strong. And that's without the China effect, by the way. That's just simply new programs, right, that are ramping at our lead customers. And then that content growth actually keeps kicking in, in '22, '23 and beyond. And you have new geographies rolling out. And we also said that the China piece, if that kicked in, that would be sort of a -- that would be a tailwind. We're not -- we're counting on China kind of being fairly flat this year. So you have that happening. You have cloud accelerating, okay, in the second half for us. And again, that's primarily based on new wins. We have automotive, right, which we had said was going to be exiting the year at about $100 million run rate, Vivek. By the way, 5 years ago, when I became CEO, that revenue was approximately 0. And we believe we have line of sight to actually achieve that run rate earlier than exiting the year. That business is going great. And obviously, automotive SAAR is up, but also we're in 24 different car OEMs, right? They're all ramping up. That's all new content for us. So that's a real tailwind. And then our enterprise business, which even if we don't get an enterprise "recovery," that has been growing well over double digits, times into the teens, purely on new product refreshes and cycles. So that all continues. And candidly, it only gets better in calendar '22. And then, let me just comment on Inphi real quick, and then I'll have Ashish at it. Inphi, we guided up -- we guided to grow up to $250 million in our Q2. The comparable period a year ago would indicate that's up about 23% year-over-year. But note that in the year ago quarter, there was also Huawei revenue in the [indiscernible]. So this was before the Huawei license requirements were put in place. So if you strip out Huawei, the Inphi business has actually been growing at about 30% a year. And that's -- and they had eSilicon in at that time. It's like a wacky one, that's apples-to-apples. So we see that business being accretive to the Marvell growth rate. And obviously, we're thrilled to see them out to shoot going to be accretive to earnings in the first quarter and also revenue well above, I think, where Street had the model, although those models were probably outdated. But even above our deal model, when we did our own diligence and our own sort of internal projections, they're tracking better. And they'll also track better in the second half than we had projected. So that all looks really good. So yes, I see an accelerating period of growth for Marvell, all of which is above the articulated new target model, which was 12% to 16% top line growth. I think, both companies can be above that growth rate, not only in the second half, but continuing into calendar '22. Ashish, do you want to add anything, if I missed...

No, I think, the only thing I'd say is, yes, I think just to be very clear, I think this acceleration in second half, that's really also, not just an implied comment, it's an organic comment, Vivek. So I think if you look at what we did in Q1, [indiscernible] Marvell up 17% year-on-year. If you look at the implied guidance, right, you take out the $250 million essentially for Inphi, then we're basically telling you it's up 17% again at the midpoint for Q2. Second half, organic, forget Inphi for a second, will be above that on a year-on-year basis. So it's a pure organic comment. Inphi's, as Matt walked through, implied 30% for Q2. They are growing faster than us, right? I mean the inside cloud data center, connectivity continues to grow. The entire 400ZR cycle, which we'll probably get to, is completely in front of us. Not just second half, but also as you go to the next couple of years. So Inphi, as we said, when we acquired them, I mean, they -- given more than half the revenue comes from cloud already, they're going to be growing faster than core Marvell did. So I think it's a great combination.

And Vivek, just one final comment. We'll move to the next one is I think one thing that's a little different as well that's happened. If you go back to some of the last few years of the churn in the business and us transitioning our business model and some of the M&A. Remember, we were fighting a number of various headwinds, various fronts on the headwind side, right? We had headwinds in Wi-Fi. We had headwinds in HDD controllers to notebooks. We had these headwinds built into the business. So it's always tougher to show that. One effect of what's gone on, besides the fact that we're in some pretty juicy, nice growth markets in the sweet spot of the market trends, we've actually managed to reduce the declining businesses in the company to a de minimis level. So we've kind of worked -- it was a labor of love to work those out. But half the battle in growth, Vivek, is not having things that are declining. It's maybe an obvious statement, but it's a big problem for chip companies. They're always big legacy businesses. They're always going down and you're always explaining it. We just tried to take the pain and get them out of there in a reasonable fashion so that there's not a lot of headwind that's in the business anymore. So I don't have like some big segment that's like, oh, yes, well, if that thing wasn't declining, we'd grow even more. We don't -- we've kind of taken that out of the equation. And so it's kind of a nice place to be finally.

Got it. Next thing, Matt, when I look at the 5G business, so Marvell is now very well recognized as a leader in 5G processors. Where are we in that 5-year deployment cycle? Because when I look back at 4G, I know it started in like 2010, and I think even until last year, telcos were adding 4G capacity. So I know these cycles can tend to last for a decade. But usually, the real money to be made is in the first big phase of deployment, right? The first handful of years. So where are you in a 5G cycle, both from just how the deployments are getting phased out in different carriers, but then I think you have this additional level of kind of share gains because you don't have much of 4G to lose, so it's all kind of incremental to you and you have new customers layering on bit right at different brands.

For sure. I mean, a couple of points. I think the first is on your point about we had very little 4G business. I mean, if you -- first of all, if you went back to stand-alone Marvell, pre-Cavium, actually if you added up 4 -- wireless, cloud and auto, right, pre all these deals, right? This number had to be -- okay, it was not -- it was like 0. Okay. So now these are the big growth drivers. So you got fast forward. So one is we do have a share gain going on between 4G and 5G, where we're gaining a lot of content. And I'll tell you this. All the money to be made for us is all in front of us, okay? When you look at these cycles, remember, phase 1 was really an FPGA-based deployment of 5G, right? It started with Korea, which was pretty robust in all the trials that went on. Our revenue started to kick in at the end of 2019, right, and through 2020. And those chips, by the way, just to give you a sense, first of all, we -- it was really with 1 customer, although we picked up some ASIC business as well with Avera. But all those products, just to give you a sense, Vivek, they're all on that sort of what I would call the 12- to 16-nanometer node, right? They're sort of TSMC 12 or 16 or GLOBALFOUNDRIES or Samsung 14. So think of that as a category that's now ramping this year. There was a whole slew of 5-nanometer awards that we got, and some of those are the basis for the public awards that were listed, for example, from people like Nokia, where those chips, and we began with one even more since then, that whole wave of sort of, call it, 5-nanometer optimized silicon to the knee of the curve on the volume, those are going to be ramping up to the '23, '24 kind of time frame. So from an innovation standpoint, Vivek, and sort of the hardware improving and the ability to deliver these base stations with the right capacity is at the right cost points, there's a lot in front of us. And from a Marvell revenue standpoint, again, we don't make FPGAs or even one of these legacy solutions. So it's all transitioning, right, to our digital silicon products. And then even our lead customer, Samsung, they're kind of end market opportunity, which is not China, it's rest of the world, that's sort of in front of them as well. So -- and plus you got multiple customers beyond those top 2 now and you've got the O-RAM vRAN things. So yes, that -- the whole thing is like we're in the very, very early innings for Marvell in terms of revenue opportunity. It will take years before it peaks.

Got it. Matt, on the storage side, right, that's -- that part of the business has kind of a mix of some legacy exposure, right, some kind of more stable markets and some new growth areas coming up. What is the right way to think about like a 3- or 4-year growth forecast, right, or at least a band of where you see storage growing over the next 3, 4 years?

Yes. I'll make some comments, and Ashish, you can chime in. And maybe you'll be sad, Vivek, but we -- this will be probably the last quarter where we'll actually report it. But I'm glad to go out on a high note, and that was sort of like the end of my speech in the earnings call. But look, the fact that the business now is growing double digits, okay, is a real testament to what the team has accomplished. I used to talk about low single-digit growth in storage and then everyone would laugh at me and say, why isn't it going to decline 10% a year like it always has. So our team has done a great job primarily because we've pivoted the market mix, right, to being more data center oriented. So 60% kind of that business now is data center levered. But then the other parts of it are things like smart video, okay, industrial drives. We've got DIY SSD, right, for kind of high-end consumer-type applications, right? So if you sort of layer the data center and then you layer the SSD and then you layer the industrial sort of smart video fit, there's not a lot left that's bad. So again, it's that headwind issue. There's not -- there's -- the hard drive notebook 2.5-inch thing is -- it's just tiny, right? There's just not that many notebooks that even ship with hard drives anymore. So that business can continue to grow. Certainly, the controller, the storage controller portion, I think it can very much continue to grow at that kind of healthy rate. The fiber channel piece, that's a flat business to sort of secularly declining slightly each year. It's high margin. So you blend those 2 things out. It's clearly better than our sort of low single-digit narrative we had before. It's going to do much better. And candidly, our internal teams got very aggressive plans there because, I mean, just look at some of our announcements as an example. We announced a key win, 5-nanometer PCIe Gen 6 data center, cloud class storage controller. Those are swim lanes we never had, okay? But that's all incremental, plus it's such a leadership product. Preamps is another example. That's in our storage bucket. That was, I guess, how much the revenue was 5 years ago? 0. Now we said it's a $50 million run rate. It's set to at least double and probably can keep going. So there's some nice growth drivers in there, right? There's cloud DIY SSD. I mean, I can kind of rattle them all off, but the team has done a great job that went and took this sort of left-for-dead business and turn it into a vibrant growing asset of the company that also, by the way, produces very, very strong cash flows, gross and operating margins is differentiated. So yes, it keeps growing.

No. The next thing, Matt, I wanted to quickly touch on is you have this really interesting asset inside the company, which is your ARM server IP, right? And I think you have been pretty straightforward that, look, will only do kind of customer programs, you're not going to build a general purpose processor. But there is growing interest in ARM servers. In fact, a few days ago, we had a chance to talk with Ampere, right? And NVIDIA is talking about launching an ARM server as well, right? Amazon has some internal programs. What do you see as kind of the monetization potential of this really interesting IP that you have? I know you use it in, right, that ARM, right, multicore or IP, in many other parts of the business. But what is the monetization potential for this asset that you have?

Yes. I think it's very high. I think that all those things you pointed out, whether it's AWS -- AWS and Graviton, I think, was the tipping point in terms of proving that this is an at-scale type of technology that can be used. We did make a business model shift last year, which basically went from us spending tens of millions or lots of money per year to keep making a standard part, which basically the entire industry took and said it's really nice. Let me run all my software prototype. Let me buy a few servers. But I mean we had a great product. We did roll -- but did anyone ever buy it? There are a lot of nice press releases. Even [indiscernible] was putting out press releases. And we were -- everyone's been so excited. So I think where the meat on the bone is, if you really look at this application, where the volume will be, will ultimately be in customized versions of this. That's what we thought last year. And my conviction has only risen but that's going to be, for us, the best go-to market. Now for others, whether you're NVIDIA who will have probably the scale to go do that, plus they have their HPC whole angle with GPUs kind of makes sense, and then they could go from there. Someone like Ampere is a very well-funded, venture-backed startup that -- maybe that business model is situated. The venture business model might be situated to kind of go explore in that regard. But we treated this asset like that internally, and we just thought it was going to be difficult to -- because nobody wants to buy the standard part. They just don't. We've talked to every potential customer on the earth 10x over, and it all looks nice. But in the end, people want their own things. So we don't have any details to disclose at this point, but I'm very confident that there's absolutely monetization for us in that area, and we'll be a participant. And there's huge leverage for Marvell, as you pointed out, because we're probably arguably the most important ARM licensee and partner on the infrastructure side. So all the work we do on the fabric, the I/O, the interconnect, all the knowledge we have to build these high-end CPUs for networking, for base stations, for enterprise, all of that's so leverageable, right, over to the server market, and we have all that history from doing actually the Thunder product line. So we're a good choice for customers to pick.

Got it. And then finally, Matt, in the last 2 or so minutes that we have left, what does Inphi mean in terms of both your kind of blended exposure to the cloud, because they have very high exposure to the cloud. What does it mean in terms of the kind of the revenue synergies with what you have on the networking silicon side? And then what does it also mean in terms of, they had a number of interesting acquisitions in making custom products, right? And you also have a lot of custom programs. So I imagine there is some synergy, right, in those kind of assets. So what does the addition of Inphi mean to the Marvell business?

Yes. Well, first, it's a great way to close. I think the first is, I've run this company from day 1 with a market-based view, not a product-based view, right? That's the whole thesis . I feel like for chip companies, market always wins. That's the place you got to focus first, right? Forget about you build the best way to, okay. The widget, on its own, probably doesn't mean anything if it's in a crowdy market. So on that -- so one of the most attractive things was their cloud exposure to your point. And it's well over well over 50%, right, of the revenues, probably more -- higher than that. And by the way, when we report our end market data, which we'll have a data center breakout, I think it will be very illuminating to see really the combining heft this company is going to have in data center. That is going to be, by far, the biggest end market in terms of revenue for Marvell starting the second we reported. Okay. Way bigger -- already there. Revenue in data center, way bigger than 5G, way bigger than enterprise, et cetera. So that's enterprise networking. So that's very exciting. Now we also -- and on top of that, the place where they fit is in a very difficult, high differentiated area, which is this electro-optics interconnect technology, which fundamentally, if you think about it, connects Marvell chips inside the data center to each other, to switches or to other servers. So it's a really logical fit for us. And it turns out there's very good customer cross-selling there with respect to, how can we do more together? And then starting discussions now on the custom side about how can we pull some of the Inphi IP into future designs. On the eSilicon side, which was a -- there was an acquisition they did of an ASIC franchise. What it actually did is it's bolstered our asset capability, right? We've literally got all that revenue now. We've combined with our ASIC, our existing Avera team. And they've taken those relationships, right, that Inphi had. And now we're going back to those customers. And guess what, we're showing them 5-nanometer, we're showing them the platform. And then we're also going to them and showing them the same model, which is, hey, not only can you get Marvell ARM CPUs, right? You can get all of our IOs, you can get, hey, we're showing a co-packaged optics roadmap now to these ASIC customers. We're actually showing a coherent DSP as an IP, right? Hey, why build your own coherent DSP? We can go give you the big bulk of it and you can use your forward error correction, unique thing that you want. We can stitch it together. So there's all kinds of unique things we have with this custom asset now that we've expanded to include IPs from Inphi as well as Marvell, and it's given us more heft in the ASIC area. So it -- so that part never really got talked about when we did the deal just because it was sort of -- would get lost in the noise. But we actually have an even more fully scaled custom offering, which I think is really well received by the customers. And then it turns out some of their ASIC customers is some of our ASIC customers. So then we can get them all on the same roadmap. And there's been a lot of benefit there. But yes, it did...

Accretive, right? Vivek, right, also...

And then, yes, if you just kind of end it, I think the one area where we got a little banged at and hammered actually at the very beginning, and of course, that was because there was a leak and all the information was wrong, but still there was a lot of consternation when this deal was announced, primarily because of the purchase price and that given the stock component that it was going to be dilutive to earnings, and it was going to take us some time to sort of get it to do that. And ultimately, I think all our investors got comfortable with that, that it was worth sort of the weight, if you will, because of the other benefits. But what's great to see is that's going to be off the table, right? We guided this thing to be earnings accretive in Q2, our first quarter out, and that's primarily because the revenue is well above what we had modeled. And it's a strong gross margin business, and we're already getting some expense rationalization. So if you do the math, it actually adds to earnings next -- this coming -- the quarter that we're in. And that's just a great thing to have sort of once we close the quarter, have that behind us. And then it's about how much earnings power can this drive over time because think about a top line growth in that 20%-plus range, right, whenever -- they were a mid-60s percent gross margin company, and we've got synergies layering in, right, over time, which will come out of us and come out of them, right, the rationalization. And the -- and part of that is the benefits of scale, right, the benefits that are accrued to them. So I think the asset can deliver a lot of earnings power to the combined company. And strategically, they're so critical to these key customers, right, that you end up just having, again, a different dialogue, like we did on 5G, where we had that baseband. It turned into, okay, well, what about the CPU, what about the switch? What about the PHYs? What about the custom ASIC for the beam forming? I think it's almost like that world is going to open up for us on the cloud side as well.

Excellent. On that positive note and that enthusiasm and that passion, let's wrap up the session there. Thank you so much, Matt. Thank you, Ashish, for joining us and sharing your insights about the industry and Marvell. We really appreciate it.