Marvell Technology, Inc. (MRVL) Q4 2023 Earnings Report, Transcript and Summary

Good afternoon, and welcome to Marvell Technology’s Fourth Quarter and Fiscal Year 2023 Earnings Conference Call. All participants will be in listen-only mode. [Operator Instructions] After today’s presentation, there will be an opportunity to ask questions. And please note that this event is being recorded today. I would now like to turn the conference over to Mr. Ashish Saran, Senior Vice President of Investor Relations. Please, go ahead, sir.

Thank you, and good afternoon, everyone. Welcome to Marvell's fourth quarter and fiscal year 2023 earnings call. Joining me today are Matt Murphy, Marvell's President and CEO; and Willem Meintjes, our new CFO. Let me remind everyone that certain comments made today include forward-looking statements, which are subject to significant risks and uncertainties that could cause our actual results to differ materially from management's current expectations. Please review the cautionary statements and risk factors contained in our earnings press release, which we filed with the SEC today and posted on our website, as well as our most recent 10-K and 10-Q filings. We do not intend to update our forward-looking statements. During our call today, we will refer to certain non-GAAP financial measures. A reconciliation between our GAAP and non-GAAP financial measure is available in the Investor Relations section of our website. With that, I'll turn the call over to Matt for his comments on our performance. Matt?

Thanks, Ashish, and good afternoon, everyone. Let me start by welcoming Willem, who is participating today in his first Marvell earnings call since being named CFO in January. Having previously served as Marvell's Chief Accounting Officer and Treasurer since 2018, Willem has deep institutional knowledge of our company and our end markets, which has helped him seamlessly transition into his new role. I look forward to partnering with him, as we continue to execute on the many opportunities in front of us. Turning to our business results for the fourth quarter of fiscal 2023. Revenue was $1.42 billion, growing 6% year-over-year, above the midpoint of guidance, with better-than-expected results from our data center end market. Sequentially, as our customers dealt with a broad-based inventory correction, revenue declined by 8%, with the majority of the reduction coming from storage products within our data center end market. The rest of our end markets held up relatively well in a worsening macroeconomic environment. Revenue grew sequentially in carrier, consumer and auto industrial and declined slightly in enterprise networking. Looking ahead to the first quarter, the inventory correction that we described last quarter has continued to impact near-term demand, along with typical seasonality for some of our products. As a result, at the midpoint of guidance, we are expecting consolidated revenue to decline by 8% sequentially and 10% year-over-year. In addition, coming out of the supply crunch, broadening inventory corrections are creating an unusual revenue mix. In the first quarter, we expect storage to decline further and inventory correction to spread to a number of additional areas I will discuss later. At the same time, we are forecasting very strong sequential growth in revenue from 5G and a number of custom ASICs, but these have gross margins well below Marvell's corporate average. As a result, we expect a challenging gross margin outlook for the next few quarters. However, we are confident that once we emerge from the inventory digestion phase into a more normalized environment, we will be well positioned for our gross margins to recover. Willem will discuss our expectations in his prepared remarks. Let me move on now to discussing our end markets, starting with data center. In our data center end market, revenue for the fourth quarter was $498 million, declining 13% year-over-year and 21% sequentially. Revenue was higher than anticipated, driven by better-than-expected results, primarily from our PAM DSPs and data center switches. As expected, our storage business was responsible for the bulk of the overall sequential decline in our data center revenue. Our results reflect the deceleration in the data center end market and the beginning of efforts by our customers to adjust their inventory to respond to a more challenging market conditions. We are expecting this trend to continue and are projecting lower demand to impact multiple data center products. As a result, we expect revenue in the first quarter from our data center end market to decline in the mid-teens sequentially on a percentage basis. We project data center storage to decline again sequentially in the first quarter across HDD, SSD and fiber channel and also expect to see inventory adjustments to a lesser degree, broadly impact the rest of our data center products. Slowdown in spending signaled this year by multiple large data center customers is also impacting the timing of the ramp of our cloud optimized design wins. Our key design win projects remain intact, but the start of production for some of these programs is being delayed. As a result, the revenue ramp has shifted out by a couple of quarters compared to prior projections. Our lifetime revenue expectations from these design wins remain in the same range as previously communicated. While we work through the near-term dynamics in the data center, we remain confident in the growth outlook for this end market. We are seeing data center customers prioritizing key growth areas such as AI and ML with potentially much larger investment over the next few years. Our relationship with Tier 1 cloud customers has continued to deepen as we engage with their architects on helping solve their most pressing challenges in their next-generation data centers with optimized customer specific solutions. One key example is the collaboration we announced with Amazon Web Services intended to enable cloud-first silicon design, extending the long-standing relationship between our companies. Marvell is a strategic supplier to AWS, delivering cloud optimized silicon that helps meet the infrastructure needs of AWS customers, including the delivery of electro-optics, networking, security, storage and custom design solutions addressing multiple critical applications. We believe Marvell's leadership in essential silicon technologies helps AWS push the boundaries of data center performance essential for driving their long-term growth. Earlier today, we announced our new Nova optical DSPs and Teralynx 10 switches for next-generation data center networks. Nova is the first commercially available PAM4 optical DSP to provide 200 gig per wavelength, twice the throughput of existing solutions over the same physical fiber. This breakthrough made possible by a significant improvement in electro-optics technology enables the industry's first 1.6 terabits per second pluggable optical modules. These modules provide twice the bandwidth in a physical package similar to existing solutions and are essential for the full deployment of 51.2T switches within the size and thermal density constraints of data centers. The higher performance enabled by this new 200 gig, 5-nanometer PAM DSP helps to extend the lead in electro-optics Inphi established and driving the market from 25-gig NRZ, 50-gig PAM to 100-gig PAM. The Nova platform provides a complete 1.6T solution, including DSPs, TIAs and drivers to the optical module ecosystem, reinforcing our expectations that pluggables will continue to remain the backbone of high-speed optical networking within data centers for the foreseeable future. Teralynx 10, a 5-nanometer programmable 51.2 terabit per second solution marks the latest in a series of cloud-optimized, low-latency, high bandwidth switches designed for use in leaf-and-spine architectures. The switch was designing close collaboration with leading cloud customers and Teralynx 10 together with Nova, creates a full platform to enable the next leap in bandwidth for cloud data centers. This is a tangible realization of the benefit of combining Marvell, Inphi and Innovium into a single entity focused on data infrastructure. This combination of 51.2T switches with 1.6T optics enables a quadrupling in bandwidth versus existing solutions. This significant breakthrough in capacity at a lower power and cost per bit will be a compelling TCO driver for customers to viably upgrade their networks to support the large increase in bandwidth they need for AI and other applications. For customers, the co-deployment of Nova-based pluggable modules and Teralynx 10 switches should reduce their risk and accelerate time-to-market. And when combined with Marvell's extensive verification and interoperability testing should also minimize their work in transitioning to a new technology. Turning to our carrier infrastructure end market. Revenue for the fourth quarter was $275 million, growing 14% year-over-year and 1% sequentially. Marvell's wireless and wired businesses drove strong year-on-year growth. We saw a strong demand for our wireless products as 5G adoption continue to expand. Our wired business benefited from carrier backbone bandwidth upgrades that accelerated during the pandemic. At Mobile World Congress, we announced our next-generation 5-nanometer OCTEON Fusion 10 baseband processor. This customizable wireless platform has been adopted by leading base station OEMs to provide comprehensive, in-line layer 1 acceleration. This new processor, along with our previously announced OCTEON 10 DPU provides a complete processing platform for 5G baseband, transport, and massive MIMO. At MWC, we continue to see strong interest from multiple customers and partners for our latest generation of 5G products for both conventional and cloud-based architectures. Moving on to our outlook for the next quarter. For the first quarter of fiscal 2024, we are expecting significant growth in our wireless business with revenue projected to increase by approximately 25% sequentially driven by 5 G deployments in multiple geographies and our customer-specific product ramps. On the other hand, after an extended period of strong growth in our wired business, we are expecting revenue to decline in the double digits sequentially on a percentage basis. As a result, for the first quarter of fiscal 2024, we expect revenue from our overall carrier end market to grow in the mid-single digits sequentially in the mid-teens year-over-year on a percentage basis. Moving on to our enterprise networking end market. Revenue for the fourth quarter was $366 million, with a strong 39% year-over-year growth, driven primarily by higher content and growing share of our merchant products, which drove the vast majority of our revenue in this end market in fiscal 2023. Sequentially, our revenue declined by 3% as we started to decrease channel inventory of our run rate merchant products, including further reductions in China. This was partially offset by growth in custom ASICs. Looking ahead to the first quarter of fiscal 2024, we are planning for additional reduction in channel and customer inventory of our merchant products and enterprise networking. However, we expect a strong ramp in custom ASICs to partially offset this decline. As a result, we project our overall enterprise networking revenue to decline in the high single-digits sequentially, while year-over-year growth is projected to remain strong in the high teens on a percentage basis. Turning to our automotive and industrial end markets. Revenue in the fourth quarter grew 25% year-over-year to $99 million. Sequential revenue growth accelerated to 18% as supply improved. Looking to the first quarter of fiscal 2024, we project continued sequential growth from our auto business to be more than offset by a decline in our industrial business. Year-over-year, we expect our auto business to continue strong growth of over 30%, offset by a decline from our industrial business. As a result for our overall auto and industrial end market, we expect revenue to decline approximately 10% sequentially and be up – flat to up slightly year-over-year. Moving on to our consumer end market. Revenue for the fourth quarter was $180 million, declining 3% year-over-year and growing 1% sequentially. Revenue from our consumer SSD controllers continue to grow, while we saw declines in legacy printer ASICs and HDD controllers. Looking ahead to the first quarter, which tends to be seasonally softer in the consumer end market, we are forecasting revenue to decline sequentially by approximately 10%. In summary, fiscal 2023 was a very strong year for Marvell, with revenue growing 33% year-over-year to $5.9 billion, well above the industry and our long-term target model. Revenue from cloud grew approximately 50% year-over-year. Annual revenue from 5G crossed over $600 million and auto crossed over $200 million, important milestones for both end markets. During fiscal 2023, the first full year following the acquisition of Innovium, we drove a significant ramp in our data center switch revenue as a combined team. The Inphi portfolio continued to fire on all cylinders, with a strong ramp of our 800-gig PAM4 DSPs and 400ZR data center interconnect products. We completed the acquisition of Tanzanite to accelerate our organic CXL development. Our enterprise networking business had a tremendous year with revenue growing 51% year-over-year, reflecting the significant share and content gains we have driven over the last two years. Through this period of rapid growth, we scaled the company in a thoughtful manner with non-GAAP OpEx growing 20% year-over-year, well below the 33% growth in revenue. Following another strong year for design wins, our opportunity funnel continues to expand with many large sockets that we believe Marvell is well positioned to win. While the broader semiconductor industry, including Marvell, is dealing with near-term headwinds, we are confident in our ability to weather this cycle and continue to execute over the long-term. We remain laser focused on capital allocation and improving efficiency. We are evaluating our customers' latest plans, assessing our level of investment across the portfolio and redirecting resources to our best opportunities. As a result, we expect to reduce our OpEx in the second half of fiscal 2024. Over the long-term, our prospects remain compelling. We have conviction in our plan, and we will continue to invest to support our strategy. As fiscal 2024 progresses, we expect the headwinds from inventory digestion will begin to subside, and mix will improve as demand patterns normalize. While storage has been impacted the most from inventory digestion, we are encouraged to see that customers have started to reduce their finished goods stock, clearing the path for a recovery. Revenue from OEM customers in China has declined to less than 10% of total company revenue. But as China reopens, we expect demand will recover and become a tailwind. In fiscal 2024, we expect data center switching and 400ZR will continue to grow, and revenue from our cloud optimized silicon programs start to layer in. In addition, we project our 5G and auto end markets will continue to grow in fiscal 2024. As a result, for overall Marvell, we expect revenue to start growing in the second quarter and gather momentum in the second half of the fiscal year. Longer-term, we are confident that our focus on data infrastructure and exposure to diversified end markets with secular growth position us well for the future. In our wireless end market, we expect content gains to layer in as 5G adoption continues worldwide. Our automotive opportunity continues to grow, and we see a path to growing revenue to over $500 million annually over the next few years. In data center, we expect to open up new revenue streams from emerging CXL and AEC opportunities. We expect generative AI to drive a massive transformation in data center architecture. We see a bigger opportunity for cloud optimized silicon for custom compute, trend we had extensively discussed over the last couple of years. In addition to compute, the level of scaling in these generative models requires a significant innovation and technology leadership in networking infrastructure to interconnect AI supercomputers. This requires ultra-high bandwidth links with low latency and sufficient reach, minimizing energy expended to move the massive amounts of data in these platforms is another important criteria. We believe these requirements are best met by high-speed optical connections. Last year, we launched the industry's first 800 gig PAM DSP and saw a huge ramp driven almost exclusively by AI applications. Our PAM DSP revenue from AI in fiscal 2023, more than quadrupled from the prior year. As AI models continue to grow in complexity, we expect that they will require more and more low-latency bandwidth. Earlier today, we announced the industry's first 1.6 terabit per second PAM platform, enabling a further doubling of bandwidth within the AI cluster. As investment in AI accelerates, we see this as a new growth engine for our electro-optics portfolio. As you've often heard me say, our employees are Marvell's greatest resource, the culture they have built is the foundation of our ongoing success. In January, Glassdoor named Marvell as one of the top 100 Best Places to Work in the US for 2023. We are in the second highest ranking among semiconductor companies. In February, we were honored to receive the Great Place to Work certification. These awards are a testament to our culture and dedication to creating a collaborative, compassionate and respectful workplace, while remaining focused on growth and execution. Thank our entire team for their contributions to Marvell to enable us to develop leading-edge essential technology that helps power the world's data infrastructure. With that, I'll turn the call over to Willem for more detail on our recent results and outlook.

Thanks, Matt, and good afternoon, everyone. Let me start with a summary of our fiscal year 2023 results. Marvell's revenue grew significantly by 33% year-on-year to a record $5.92 billion. GAAP gross margin was 50.5% and GAAP loss per diluted share was $0.19. Our non-GAAP gross margin was 64.5%. Our GAAP operating margin was 4%. Non-GAAP operating margin expanded to 35.5%. Non-GAAP earnings per diluted share grew 35% year-on-year to $2.12. We returned $319 million to shareholders through dividends and buybacks. Moving on to our financial results for the fourth quarter. Revenue in the fourth quarter was $1.419 billion, exceeding the midpoint of our guidance, growing 6% year-over-year and declining 8% sequentially. Data center was our largest end market driving 35% of our total revenue. Enterprise networking was the next largest, with 26% of total revenue, followed by carrier infrastructure at 19%, consumer at 13% and auto industrial at 7%. GAAP gross margin was 47.5%. Non-GAAP gross margin was 63.5%, 50 basis points below forecast, primarily due to a change in revenue mix within certain end markets compared to our guidance. GAAP operating expenses were $650 million, including share-based compensation, amortization of acquired intangible assets, legal settlements and acquisition related costs. Non-GAAP operating expenses were $431 million. GAAP operating margin was 1.6%. Non-GAAP operating margin was 33.1%. For the fourth quarter, GAAP loss per diluted share was $0.02. Non-GAAP income per diluted share was $0.46 at the midpoint of guidance. Now turning to our balance sheet and cash flow. During the quarter, cash flow from operations was $352 million. This included $56 million in payments for long-term capacity agreements. In fiscal 2023, payments for long-term capacity totaled $252 million. Looking ahead to fiscal 2024, we are currently anticipating significantly lower payments for capacity and a much reduced headwind to operating cash flow. Inventory at the end of the fourth quarter was $1.07 billion, growing by $111 million sequentially. As we indicated in our prior call, we expected inventory to grow in the fourth quarter to support upcoming product ramps. Looking ahead to the first quarter, we expect inventory to start to come down and be a tailwind to operating cash flow over the year. We returned $51 million to shareholders through cash dividends. As of the end of the fourth fiscal quarter, our cash and cash equivalents were $911 million, growing by $188 million from the prior quarter. Our total debt was $4.5 billion. Our gross debt-to-EBITDA ratio was 1.87 times and net debt-to-EBITDA ratio was 1.49 times. $500 million of our total debt is due in June 2023. We are currently planning on paying this off using our cash balance and free cash flow, which we expect to improve our gross leverage. Turning to our guidance for the first quarter of fiscal 2024. We are forecasting revenue to be in the range of $1 3 billion, plus or minus 5%. We expect our GAAP gross margin will be in the range of 44.1% to 46.1%. We project our non-GAAP gross margin will be approximately 60%. This guidance takes into account the adverse revenue mix we are expecting in the first quarter, and we currently expect mix will remain challenging for a few quarters. Once we get through this period of inventory correction, combined with cost improvement actions underway, we expect non-GAAP gross margin to get closer to the low end of our target range by the fourth quarter. We project our GAAP operating expenses to be approximately $687 million. We anticipate our non-GAAP operating expenses to be approximately $460 million. This forecast includes a step-up from the prior quarter due to typical seasonality in payroll taxes and employee salary merit increases. As Matt discussed, we intend to continue to invest prudently in our long-term growth initiatives while further improving efficiency and optimizing capital deployment. Altogether, we expect to reduce our non-GAAP operating expense run rate by approximately $15 million a quarter in the second half of the year. We expect the full reduction to be achieved in the fourth quarter. Other income and expense, including interest on our debt is expected to be approximately $49 million. For the first quarter, we expect a non-GAAP tax rate of 7%. We expect our basic weighted average shares outstanding will be 858 million, and our diluted weighted average shares outstanding will be 863 million. As a result, we anticipate GAAP loss per share in the range of $0.12 to $0.20 per share. We expect non-GAAP income per diluted share in the range of $0.24 to $0.34. Operator, please open the line and announce Q&A instructions. Thank you.

We will now begin the question-and-answer session. [Operator Instructions] And our first question will come from Ross Seymore with Deutsche Bank. Please go ahead.

Hi guys. Thanks for letting me ask a question. Matt, I just wanted to ask about the general business conditions. In the last quarter, you talked about -- in the October quarter, you said that things were weakening towards the end of the quarter and then throughout the January quarter. Can you talk about the linearity of demand, whether it be by end markets and/or geographies as you go through the end of the January quarter and thus far into the April quarter?

Yes. Thanks Ross for the question. We've seen the business conditions continue to weaken, I'd say, in the last couple of months, the I'd say the breadth and some of the steepness of the decline. We've continued to see -- we had thought that storage was going to kind of bottom out. Last quarter, it's down again. There are some green shoots there just given what we see in terms of end customer stock levels coming down, and we believe that throughout the year, that's going to improve. But I'd say just, generally, in the storage area broadly and then in the data center, as you can see from our guidance. not only in storage, but a little bit broader. The inventory correction is still going on. At the same time, I think what's happened in the wireless business in 5G, that's continued to be very resilient and growing, which is a positive as automotive as well continues to be very strong for us, and we see that as a growing business again this year. That's on a great upward trajectory. So, a lot of moving pieces, Ross, but I'd say the things that were weaker a quarter ago have continued to weaken a little bit more. And then the areas where we had some optimism, those have continued to do better. It's really interesting how these various end markets are kind of moving at different times. But when you add it all up, basically, and you look out over the next few quarters, we do see growth in the second quarter and then as a result of really what we believe will be the inventory digestion and the normalization in the back half plus continued strength in things like 5G and automotive and then also our cloud optimized ramp. And in theory, some of the China business should come back, we're starting to be optimistic about the second half relative to the recovery. So that's some of the moving pieces, and I agree, it's -- in the infrastructure business, this doesn't move up and down typically as much as, say, a consumer business. So it started in October. We're still going through the journey here in the January and April quarters, and then we see a recovery in the second half.

Thank you.

And our next question will come from Vivek Arya with Bank of America. Please go ahead.

Thanks for taking my question. Matt, I'm curious what second half recovery looks like? Is there the potential to get back to this kind of $1.5-ish billion quarterly run rate, you guys were at. And then as part of that, I'm also curious to know what should be the updated view $400 million and $800 million in cloud-optimized silicon. And the one thing that I think Willem mentioned that you will only be able to get to the lower end of the gross margin range. I'm a little confused that if there is the recovery, then why would you be at the lower end of the gross margin range. So kind of related questions. Thank you.

Sure. Yes. No problem, Vivek. We'll treat it as one integrated question. So let me start with the cloud optimized ramp. As I said in my prepared remarks, and you can see sort of what's happening in the data center area, there is some delay, and there's some prioritization within those companies on how they're going to deploy. We feel very good overall about the programs and the lifetime revenue of those programs, but they have, in aggregate, shifted out by a couple of quarters. So if you just assume that the plan for $400 million this year was more back-end loaded as those ramped then the effect of that pushing out by a couple of quarters, we think it probably takes it down by about half. We don't know quite yet exactly, but just to give you a big round numbers, I think that's a safe assumption for now. The other question was around what does the second half recovery look like. I think it's too early to call it specifically at this point. I mean we're still seeing a lot of volatility in the business, as you can see from our first quarter guide. But when we step back and look at it from 30,000 feet and we also take the broad indicators from our customers, we do expect it as we said, to start growing from Q2 and beyond. The question is how big is the recovery and what comes back. So it's a little bit hard to call precisely where that's going to be at this juncture. And then on gross margins, again, it really comes down to the mix. Historically, if you look, we've -- the company has actually had pretty stable gross margins. I mean, they've moved around a little bit, but in general, that's been a fairly predictable thing. You can see based on this downturn we're going through just the volatility that's come in relative to the mix. So we see that improving in the second half. Again, I think it's too early to call the exact trajectory. So just in the abundance of being prudent here, we're calling it getting up to the lower end of the range. But again, it's highly dependent on mix, product ramps and what the end markets do. This is the best we know at this time.

Thank you, Matt.

Our next question will come from Toshiya Hari with Goldman Sachs. Please go ahead.

Hi, good afternoon. Thanks so much for taking the question. Matt, I wanted to ask about generative AI and to the extent your customer conversations have evolved over the past couple of months. You talked about the cloud optimized opportunity being pushed out. But have you sensed any change in customer pull as it relates to AI? You talked about your exposure in compute and networking. You also talked about your PAM DSP business, specifically in AI quadrupling last year. Curious how you're thinking about that business in fiscal 2024? Thank you.

Sure. Yeah. I think the -- as we said in the prepared remarks, the impact of generative AI and the buzz around that is a real positive for Marvell. The ramp on our 800-gig products really over the last few years was driven by AI clusters and AI applications. So that's gone very, very well, and we see a tailwind there over time on our optics business. I'd also note, we just announced today our 1.6 terabit per second product, which also has native 200-gig per lane I/O. That's a real product that's sampling, and it's going to go into production, we believe, sometime next year. And again, the big pull there will be because of the doubling of the bandwidth, another ramp for us relative to AI. So that under the hood, despite a lot of the chop in the other businesses, has been a real bright spot for us. And we're very encouraged on that outlook. And then, in addition, some of these systems for AI clusters are going to get -- have to get -- data centers are going to have to get rearchitected quite frankly. And we think that's also going to drive the need for cloud optimized silicon as well to complement the GPUs that are used, as well as the optics. So, I think, there's a broader opportunity here for sure. I think this is a very bright spot for us in the cloud. And I think it's evolving very quickly. Ashish, did you want to add anything to that?

No, I think the other thing I'd add is, I think we're already starting to see requests from cloud customers to really accelerate the availability of our next-gen products, right? I think we've seen a distinct change in tone, I would say, over the last quarter or two where we introduced our 1.60 as an example. You're certainly getting a request at a, how quickly can it become available, because they see the bandwidth needs really expanding very, very quickly. So I think it's very critical for us to get these products out even faster.

Thanks, guys.

And our next question will come from C. J. Muse of Evercore. Please, go ahead.

Yes. Good afternoon. Thank you for taking the question. Just routine part for you, Matt. So I won't take too long. But specific to data center, I was hoping we could drill a little bit deeper into kind of the trends you've seen year-on-year based on the implied guide. So roughly, I think, down to $220 million. Is there a way to kind of parse through how much of that is storage, preamp, fiber channel and other? And maybe help us better understand kind of the slowdown in other, what drove that? And then from here, how do you think about the different parts between kind of storage, other and the new products that you already kind of spoke about. So we have a pretty good idea there. But any way to kind of frame that would be very helpful.

Yes. Sure. Yes, C. J., sure, go ahead. I'll let Willem start off, and then I can add to it as appropriate. Great question.

Yes. So, I think, first of all, on a year-over-year basis, storage is a much bigger part of that decline for Q1. The way we think about it, C. J., we spoke originally about the $1.4 billion for storage, right? And so, if you look at that on a quarterly basis, it's about 60% was data center. So on a quarterly basis, it's just over $200 million. And last quarter, we obviously indicated that was down significantly. And this quarter, again, the decrease in data center was more in storage compared to the rest.

I guess, is there a way to frame what the other parties outside of storage? I mean, it certainly feels like roughly maybe $150 million out of the $220 million decline in storage, but we'd love to isolate on that 70 and how you see that part of your business recovery?

Hey CJ, it's Ashish. Let me just add some more color. So I think if you look at the year-on-year, the much bigger portion of storage, as Willem just walked through, if you think about all the other product lines, right, this include optics as well as some of our networking products within data center, that's where we're seeing a much smaller decline in Q1. These inventory corrections take typically a couple of quarters, right? So I think you'll expect them to start to bounce back, I would say, in the second half. And I think they get a lot bigger. So the non-storage portion of data center really starts to recover in the second half and gets a lot bigger. And on top of that, even though we have pushed out some of the ramps of our cloud optimized, on an incremental basis, you will see some additional revenue in the second half. So overall, non-storage is a big part of how we see data center recovering in the second half of the year.

Thank you.

And our next question will come from Joe Moore with Morgan Stanley. Please go ahead.

Great. Thank you. I wonder if you could talk about the gross margin pressure in the April quarter, carrier being up mid-single digit is not that much incremental revenue. Can you just give us some sense for the gross margin disparity between the carrier centric custom business and the rest of it? And then I guess, separate from that, as you ramp other parts of custom ASIC into cloud, do you expect to see that be more like the corporate gross margin.

Yeah. So let me start. So the way to think about that is, I think in the prepared remarks we just said wire was done significantly about 20% and that was offset by the growth in wireless. So net-net, you see a small growth there, but it's those two dynamics that offsetting each other.

Okay. And I mean, it seems like still even 25% growth in 5G, to cause a four-point margin disparity overall, it seems like…

Yeah, sure. So I think the way to look at it is overall, we saw additional weakness in storage, particularly in HCV and fiber channel, which typically is higher gross margin product for us. In addition, we've seen a decrease in our merchant enterprise networking business, which is also a higher gross margin product. And so that with the growth in 5G, and then also we've seen some strong increase in our ASIC business, which both of those are typically slightly lower gross margin product. And that in combination with -- certainly with the top line coming down, we've had some headwinds from fixed cost absorption. And so overall, that's driving the decrease that you're seeing.

Hey Joe, let me add a couple of things because it's a very good question. So first, fully agree with everything Willem said in the way he characterized it, which is fundamentally that we had some accretive to gross margin product lines declined fairly significantly and fairly rapidly. And at the same time, we're seeing strong growth in product lines that are less than the corporate average, and it's moved the gross margin much more significantly than anything we've experienced in some time. And maybe just to take it from the top on a bigger picture for a second. Our custom business came from the purchase of Avera back in 2019. And we knew at that time that was going to be a lower gross margin business and that was fine. And it's actually grown really well. The team has executed well. It's been a great growth driver for us. And we've been able to -- and in fact, if you look, it's probably grown faster than the overall Marvell portfolio. But at the same time, call it, from 2020 to now, we've actually been able to increase and drive Marvell's gross margins up over that timeframe. So, we've had a balanced portfolio, different product lines and diversified product lines that have a wide range of gross margins that blend to something that's been very, very healthy and in general, been very predictable. What was -- the change was really the steepness and the breadth of the inventory correction we're dealing with. And so you have these growth drivers that are still kicking in on ASIC. And then also carrier, as you point out, has always been a lower gross margin business for us. We've been very transparent about that. Those are actually continuing to grow through this cycle quite significantly. And then again, our highest margin product lines for various reasons, whether it's channel inventory correction or just customers doing their own inventory burn, those have come down a lot. So, that's why we think and we look at the numbers, over the next few quarters, that should start to normalize again. So, I hope that's helpful to provide both the detail and then the bigger picture, Joe.

Thank you.

Yes.

And our next question will come from Matt Ramsay with Cowen. Please go ahead.

Thank you very guys. Good afternoon. Matt, you guys talked about maybe the change in timing and the push out a bit of the cloud-optimized $400 million for this year. But when you look into next year, that $800 million that you guys have talked about for a while is obviously not just one customer or one program, right? And all of these hyperscale folks are going through different architecture changes, periods of digestion and you can -- there's a number of words probably to characterize it. So, I guess what I'm trying to get at is if things are delayed a couple of quarters and the early ramp of early programs, how are you thinking about that $800 million next year? Are there some programs that are pushed a bit, some that are still on time? And maybe you could just -- maybe walk through some of that dynamic? And then just a real quick clarification on a question that Joe asked. As we ramp these cloud optimized solutions, are those going to be accretive to gross margin specifically? Maybe you could address that a bit. Thank you.

Sure. So, for next year, I'd say my high-level answer is it's too early to call. And you kind of nailed it. I mean some of the programs are tracking as we thought, some even might be a little bit ahead. Some have pushed out more than we thought. So, the net effect for this year, we're just trying to call it as we see it today on the shift. Now, for next year, I don't know that it just keeps sliding. I think some of these are going to ramp at their own pace. And quite frankly, it's too early to call it and to understand what's going to happen four or five quarters from now, given how much is changing in the near term. I guess I'm reticent to try to provide any more precision there. I mean, I think if you want to be conservative, you could just sort of keep shifting it out a little bit, but I think there's still a call option for growth for Marvell next year in this area, but we have to see how these programs play out. And I think on the question of the margins, it really depends on the program. They're not all the same and some are more custom in nature that would be more traditional profile. Some are higher. But we still see the blend for the overall company returning back to its towards the low end of the range this year, and then we're going to keep driving it towards our long-term model in the next year. So at this point, I think it's too early to call and kind of infer what that the gross margins of those programs would be and then when they ramp versus what's going to happen next year. So as we get through the year, I think, Matt, we can give you more precision on that. But we're not changing our long-term growth model at this point in terms of the financial model, whether it's gross margins or long-term revenue growth and operating margins.

Thanks, Matt. Appreciate the color.

Our next question comes from Tore Svanberg with Stifel. Please go ahead.

Yes, good afternoon. This is Jeremy calling in for Tore. I guess I want to focus in a little bit on the custom silicon. It sounds like this business has proven more resilient than other areas of the data center. Is this a function of the nature of those programs and the NRE investments in your customers that may and maybe you get better visibility in true end demand? And also, is this resiliency reflected in some of your ongoing current custom programs, maybe more specifically comparing the pushouts here relative to rest of the data center business?

Sure. Yes, a couple of things that are going on. I think as I said, one, because that business and team has done really well in terms of overall ASIC growth, those programs tend to still be ramping. So there's not a lot of inventory that's been built that's been part of the growth. So that would explain some of it. But I'd also note that the breadth of the engagements we have in this traditional custom area, our data center, but they're also in carrier and they're also in enterprise. And so as Willem indicated, even within the enterprise segment, we've had some mix shift there just with the traditional merchant products, which are much higher gross margin coming down and then some of the ASIC stuff ramping up. So again, I think this is all going to get normalized over time. But certainly, because of the growth in that area, it's proven to be very resilient. And I think some of it is maybe because sure, we've gotten NRE. You get -- I think there's just -- historically, that team has had a -- those -- that type of business has been one that's been, I think, a little bit easier to plan. It seems to be a little bit more predictable, I don't know. But I think in general, it's mostly because they're new programs ramping.

And can you talk about the pushouts here relative to the other segments? Have there been any significant changes there?

Sorry, say that again?

I guess, has the custom program has been impacted by the pushouts you're seeing in other areas of data center?

No. I mean, again, I wouldn't say that's a major factor. I mean, again, some of the cloud optimized programs that we talked about, those are custom. But in the short term, the data center impact has really been, again, more driven by storage and the impact on the rest of the portfolio, whether it be custom or optics, or whatever hasn't been as pronounced, but it's still going through its own correction.

Great. Thank you.

Our next question will come from Blayne Curtis with Barclays. Please, go ahead.

Hey, thanks for taking my question. I just wanted to go back to -- I just want to understand what your message is on data centers. So the hard drive, it's more of a correction -- inventory correction at your customer. What are you seeing from just overall data center? Because I thought the message last quarter was the PAM business and switching were actually kind of okay. So, I guess, when you're talking about cloud optimized pushing out and you said data center is softer, I'm trying to understand how those two relate, I guess. And then you mentioned architecture changes, if you can just elaborate a little more. I'm still kind of confused what you're saying on data center and why that's delaying the cloud ramp.

Sure. Maybe, I'll break it into two pieces. So let's go back to -- because I think there's two separate issues we're talking about. So last quarter, when we signaled the weakness in data center, it was very pronounced in storage. And we also said we were starting to see signs of inventory correction on the rest of the portfolio, but it was more muted. And at that time, our expectation was that storage most likely was going to at least flat now, maybe go down a little bit more. And then there'll be, again, a continued more mild correction. I think what's happened is, the data center -- the storage in data centers continue to need more time to correct. So that's declined again in the Q1 guide. And then, I would say, the rest of the portfolio has also -- is now seeing more of a correction needed than we thought last quarter. Okay. So that's in the short term right now, Blayne. So, yes, storage, down a little bit more and then the rest of the portfolio, including those other product lines you mentioned, in aggregate. I mean, within underneath, some of them are moving up and down, but the net effect is that there is an inventory correction going on in data center across the portfolio. Independent of that, we -- is a separate issue. We were -- we have been anticipating a ramp of our cloud optimized design wins, which have been gathered over the last few years. That's still going to happen this year, but it's been -- some of it's been pushed out by a couple of quarters. So we still see that kicking in at the end of the year and adding to growth. But that's sort of separate than any inventory correction. That's just program execution by our customers, timing of their ramps, et cetera. And then on the third point, which is just, all we were pointing out on the architecture stuff was that, as we see the growth in AI-based systems and AI-based clusters and AI-centric data centers, we think longer term, I mean, that's now outside the window we're talking about here, whether it's this year or next year, that, that's going to be a significant growth trend and that's going to require both cloud-optimized silicon as well as our high-speed optics. So I think three pieces to it, short term, medium term and long term.

Thanks a lot.

Very dynamic overall, Blayne, I guess, that's the punch line.

Yes. Thanks.

Yes.

And our next question will come from Christopher Rolland with Susquehanna. Please, go ahead.

Hey, guys. Thanks for the question. I guess, my first is the $400 million to $800 million regardless of the pushouts. Do you have any visibility into the composition, or can you elaborate on the composition of those wins around maybe percentage storage versus network versus compute and maybe even optics in there as well, anything else that you can offer? Just any visibility into the composition would be great.

Yeah. Hey Chris, there's a couple of challenges there. So one is there's a very high level of customer sensitivity on the nature of these programs. These are all NDA based. They don't really want us discussing those. Ideally, over time, maybe we can get something more public out there, but until my customers give you the go ahead, I'm not going to do it. All I'd say is just in general, maybe to try to be helpful; it's really not much storage in there. This is really more custom compute, custom networking, accelerators, offload that type of thing. That would be the bulk of it, I would say, those types of applications.

Okay.

And as we go along, Chris, I think we get closer to these product ramps and things, I'm obviously hoping for our investors, we can continue to provide more transparency as these things materialize, but it's a bit early to talk about that mix. And I'd prefer to do that as we got closer

Sure. If you didn't like that question, you probably will like this one.

I would love your questions, Chris. Go ahead.

In terms of AI, you talked about networking, you talked about optics there and that all makes sense. But I was wondering if you had any ability to also provide compute. At one point on your road map, you had AI cards. I think you kind of tabled that effort. But yeah, I would love to know if you can provide compute into AI? And then the second part of that question is, are you willing to guide beyond the $800 million? Do you have visibility into, let's say, beyond $1 billion as it's been some time since you gave that $400 million to $800 million number?

Okay. Maybe just real high-level answer to both of those, I think on the AI one, I think that is and will be an opportunity for custom and for cloud-optimized applications. I think that's a real thing, meaning people will want to use things that are merchant. They want to use some of their own special things and probably in combination. So I think that's an opportunity that's part of the potential. And then as far as the long-term peak revenue from those, the wins, $400 million to $800 million, yeah, I think we said when we won them, it would keep going because typically, you don't hit peak until many years into the ramp. So with the design wins intact, the $800 million becomes larger over time. How big that is? I think it's too early to call. But certainly, it would exceed the $800 million on an annual basis just because really, that was our view at the time of the year one, year two ramp would look something like that.

Awesome. Thank you, Matt.

Yeah.

And our next question will come from Karl Ackerman with BNP. Please go ahead.

Yes. Thank you. I wanted to delve into data center, if I could, for a second there. It sounds like some of the softness relates to rebalancing of optical modules at hyperscalers, and that's just -- that's not a -- that's more of an industry comment, not just a company specific comment. But I guess for Marvell, does that bleed into the July quarter as well? And then bigger picture, how are you thinking about the 400-gig upgrade cycle for 2023 despite some of the near-term challenges. Does that influence your bigger picture view in terms of 2023 and then 2024, given the broad product portfolio that you have to address both 400 gig and 800 gig? Thank you.

Yes, I think at this point, Karl, you're right. I mean we -- part of the inventory digestion we're seeing is our PAM products that sell to the optical module guys, and that's a business where, again, we're kind of one step back in the supply chain. So, we would see that and we are seeing that and feeling it. When exactly that recovers, which quarter, it's going to happen in the next few quarters. The question is next -- this quarter, next quarter or the quarter after, but it's in that probably that timeframe because the underlying demand is still strong. And then I think your other question was on -- and then, of course, the headwind we face now becomes a tailwind in the second half, if that's the case. I think your other question, was it on the 400-gig ZR or was it on the 400-gig inside data center PAM--?

Your broader portfolio on 400-gig and 800-gig and I guess, how you think about the upgrade cycle because obviously, softer the market near-term, but how do you think about your portfolio broadly as you think about that upgrade cycle over the next couple of quarters? Thanks.

Yes, I mean, we tend to think of the upgrade cycle as quite frankly a multiyear rollout, right? We had a very strong year last year, and it's a blend of products, right, from 200, 400, 800 inside data center, 400 ZR between data centers and that 1.6T announced. So, I'd say that the demand and the shift to PAM continues at a very strong rate. Again, we have some inventory correction in the first half. That's more of a module-related issue. And again, the slope of the hyperscale cloud guys revenue growth or CapEx growth kind of dropping, but not declining just not at the same rate of acceleration. So, we think all that works its way through and end consumption continues to grow very strongly last year, this year, and the year after.

Thank you.

And our next question will come from Quinn Bolton with Needham & Company. Please go ahead.

Hey Matt. I guess one clarification and then one question. The clarification, just you talked about the growth in -- driven by AI clusters for high-speed optics. Do you guys -- are you impartial to whether those AI clusters are connected through Ethernet or InfiniBand because I think different hyperscalers have different fabrics for those AI clusters? And then my question is, you guys have talked about the growth in PAM. We've got the OFC show next week, and I think there's more talk now about the linear drive modules where you don't need the DSP or the DSPs and other parts of the line card. I'm just kind of curious if you could address -- do you see any threat from direct driver, linear drive modules going forward to that PAM4 business? Thanks.

Sure. Yes. No. So, it's an exciting time for optics at this point given we've got OFC coming up next week. Yes, on the first one, in general, we're agnostic to both of those. We tend to just partner closely with the customers and design the solutions they need. So, it's the fundamental IP that really matters. As it relates to the disruptive technologies that are lurking around out there, I think that's been the history of optics. Our view is pluggables is going to be the high-volume de facto standard for many years to come. And certainly, we're encouraged by showing demonstrations of our new product at 1.6T, you had OFC, there's real people building systems around this. And of course, the ramp last year of 800 gig gives us confidence under the ramp of 1.6T. So we're paying attention to all of it, Quinn, I would say. We've got a really excellent team that understands the trade-offs and understands the various architectures. But right now, we feel very confident in our approach and our solution set. And I think coupled with our switch platform, which has now also been announced, it's really a nice combination of the assets of Marvell plus Inphi plus Innovium, all coming together in a single platform. So more to come on that. And certainly, we're happy to debrief with you and the rest of the team at and after OFC. I think there's going to be a lot of activity there.

Thank you, Matt.

And our next question will come from Srini Pajjuri with Raymond James. Please go ahead.

Thank you. Hi, Matt. Question on your wireless business. Obviously, it's good to see that business doing quite well. Some of your peers have reported some weakness in that market. Just curious as to what's driving strength for you in particular? And then as you look out to the next few quarters, can you -- I mean, to the extent that you have visibility, can you talk about the sustainability of that business? Because every time we see 20%, 25% growth in any semiconductor business, we do start to worry about potential inventory build. So if you can talk about what you're seeing out there, that will be helpful. Thank you.

Sure. Yes. No, I think Srini, I think the big issue, especially with maybe some of the more larger incumbent players that have large carrier communications businesses, especially in the wireless area, they tend to be -- have a combination of some legacy, let's call it, like 4G LTE type of solutions, and then they've got new 5G chips that are ramping up, and then there's some mix of the two. But as you can see with carrier spending, the bulk of it in wireless is just pretty much all going to 5G at this point. And so from a Marvell perspective, we had very little legacy before on the older standard and 5G is really where all the content gains rolled in. So that's continued to be a nice growth driver for us as a result. And I'd say on top of that, you've been around the block enough to see this as well, I mean carrier tends to be a very lumpy business. So it doesn't really historically ever move in a very linear fashion that tends to have be a little bursty at times. And certainly, we're happy with the growth we're seeing, and we think that that's -- overall, this year is going to be a growth year for us in 5G. But certainly, it can be a little bit lumpy. But I would say that this is not something that overall is kind of going according to plan in terms of the designs we won, what the content we could capture would be. And finally, if you just look and we said it in the prepared remarks, last year, we crossed the $600 million mark in 5G, which was a bogey we had set several years ago. We haven’t time bounded by the way, like it's going to be in this exact year, this exact four quarters. We had left it a little bit open, but we're pretty happy that really just a few years after we talked about that goal, we were able to achieve it last year. So yes, there's some bright spots for sure, in the Marvell portfolio. 5G is one of them and obviously, automotives and other, and there's some good things going on. We just have to work through the issues in some of the other end markets. And with that, I think, Ashish, is that our last question?

Yes, it is, Matt. If you can maybe end with some closing remarks, that would be great.

Yes. Perfect. Well, I appreciate everybody's time on the call today. When I look back at our last fiscal year, it was a great success for Marvell. I think it -- and I was reflecting back, looking at where we were back, let's say, in 2020, during the pandemic, we were in the process of combining with Inphi and I sort of look a few years later, and I think in fiscal 2023, we -- ourselves, Inphi and also, by the way, the Innovium team we brought in, really everybody executed extremely well. It was very, very strong growth, very strong margins, great growth overall. Certainly, we're going through, now after several years of extremely strong growth, a correction in inventory, that's more pronounced in some markets than others. We're managing through that very prudently and expeditiously and doing that also in a partnership model with our customers, to where we're trying to understand their real demand and get back to a point where we're shipping back to real demand. At this point, we believe we're shipping below consumption of Marvell products. And so, we want to get back there. And when we do that, it will also lead to the increase in recovery back in the gross margin profile. And then we can have all of our new wins that we've gotten and all the hard work, our sales team, business unit team has done over the last few years, driving our design win funnel to recognize that growth. So, I'd say, despite the short-term issues, we think very long term at Marvell, okay? I mean, I'm entering my seventh year as -- actually, about to complete my seventh year as CEO. The prospects for this company are tremendous. We have conviction in our end markets and the long-term opportunity. And the R&D investment that's needed to really execute the programs our customers want to is right in front of us. We have an opportunity, I think, to really reassess and make sure that we're focused and laser focused on the best opportunities. I think we've made capital allocation from day one, a strategic priority for Marvell. And I believe, we can come through this cycle much -- as a much stronger company with our growth prospects intact and actually have a better long-term opportunity doing really key programs for our customers with the best ROI. So we're going through that whole process now, as we do always, and I feel very good about the future of the company. So appreciate the time today, everybody, and look forward to catching up post earnings.

The conference has now concluded. Thank you very much for attending today's presentation. You may now disconnect your lines.