Earnings Call Transcript
Marvell Technology, Inc. (MRVL)
Earnings Call Transcript - MRVL Q2 2024
Operator, Operator
Good afternoon, and welcome to the Marvell Technology Incorporated Second Quarter of Fiscal Year 2024 Earnings Conference Call. All participants will be in listen-only mode. After today's presentation, there will be an opportunity to ask questions. Please note, this event is being recorded. I would now like to turn this conference over to Mr. Ashish Saran, Senior Vice President of Investor Relations. Please go ahead.
Ashish Saran, Senior Vice President of Investor Relations
Thank you, and good afternoon, everyone. Welcome to Marvell's second fiscal quarter 2024 earnings call. Joining me today are Matt Murphy, Marvell's Chairman and CEO; and Willem Meintjes, our 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 measures is available in the Investor Relations section of our website. Before I turn the call over to Matt for his comments on our performance, let me highlight several new product announcements, starting with our electro-optics portfolio. First, we announced the release of Orion, the industry's first 800-gig coherent optical DSP for pluggable modules. Orion is our fifth generation of coherent DSPs and is produced in 5-nanometer technology incorporating our 112-gig SerDes, enabling 800-gigabits per second of throughput within the tight power and space constraints of small form factor pluggable optical modules. Orion enables high-performance probabilistic traffic shaping in addition to supporting standards-compliant transmission modes. We expect that Orion is well-positioned to continue to drive Marvell's leadership in coherent technology in both the carrier optical transport market as well as data center interconnects, or DCI, in the cloud market. In parallel with Orion, we launched the industry's first 800-gig DCI ZR module, our COLORZ 800 platform. We will be demonstrating our 800ZR module at the ECOC Conference in October. Marvell pioneered DCI technology at 100-gig, followed by a 400ZR platform, which has been shipping in volume since last year. COLORZ 800 will be our third generation of DCI modules. Powered by our new Orion DSP, COLORZ 800 incorporates Marvell's innovative silicon photonics technology, which integrates multiple discrete components on a single die. Marvell's 800ZR modules provide twice the bandwidth of current solutions while lowering power and cost per bit by 30%. Our 800ZR modules will enable the deployment of next-generation 51.2T switches and routers by cloud operators to support the massive increase in traffic between data centers, driven by continuing growth from Generative AI. Moving to our copper connectivity portfolio, we announced the industry's first 5-nanometer multi-gigabit PHY platform. This platform represents a significant leap in performance compared to products on the market today and is a key milestone in Marvell's journey to physical layer technology leadership. It is based on an innovative architecture that includes optimized circuit design, custom digital logic, and enhanced DSP algorithms. This PHY platform will deliver 10-gig performance at half the power of previous generation Marvell devices and will become the building blocks for multiple standalone PHY products, integrated SoCs, and custom ASICs optimized for specific markets and applications. We expect the adoption of multi-gig PHY will continue to grow in enterprise networking. In our automotive end market, we announced the industry's highest capacity automotive central Ethernet switch to support the zonal networking architectures of next-generation vehicles. This addition to our Brightlane family of auto Ethernet switches delivers 90-gigabits per second of bandwidth, nearly 2 times higher than current commercially available solutions. The new switch family also includes a combination of advanced security features not found together in any other automotive switch product. These features include MACsec link security on every port, deep packet inspection for heightened intrusion detection, and an embedded hardware security module for encryption. This product has started sampling to multiple leading automotive customers and partners. Let me now turn the call over to Matt for his comments on the quarter.
Matt Murphy, Chairman and CEO
Thanks, Ashish, and good afternoon, everyone. For the second quarter of fiscal 2024, the Marvell team continued to execute, delivering revenue of $1.34 billion. These results were above the midpoint of our guidance, primarily driven by demand from AI applications growing faster than our prior forecast. Our non-GAAP operating expenses were better than guidance due to an acceleration of the cost reduction plan we outlined last quarter. As a result, our non-GAAP earnings per share was $0.33, $0.01 above the midpoint of our guidance. We are pleased with our performance for the quarter in a challenging macro environment. Let me now move on to reviewing our results and expectations by end market, starting with data center. In our data center end market, revenue for the second quarter was $460 million, growing 6% sequentially, well above our guidance for a flat outlook. We were able to outperform our guidance in this end market because of accelerating demand for optical products to meet the continuing expansion of cloud AI deployments. Our overall revenue from cloud grew over 20% sequentially. Notably, revenue from both cloud AI and standard cloud infrastructure grew sequentially, with AI growing faster. As expected, revenue from the enterprise on-premise portion of our data center end market declined significantly on a sequential basis in the second quarter, reflecting a weakening enterprise market. As you heard in detail last quarter, AI infrastructure requires a staggering amount of high-bandwidth connectivity, best provided by an optically connected infrastructure operating at the highest available speeds. Marvell is enabling AI with a broad range of solutions, which include: PAM4-based optical DSPs and AECs for connecting accelerator clusters inside AI data centers; DCI products for connectivity between regional data centers; low-latency high-capacity Ethernet switches for fabric connectivity inside data centers; and custom silicon for compute acceleration. We are confident that the breadth of Marvell's technology positions us as one of a scarce few semiconductor companies that can enable the industry to capitalize on the rapid growth in AI. Marvell's market-leading PAM4 optical DSPs are indispensable for the pluggable optical module ecosystem that cloud customers rely upon to build their massively scalable networks. Our DSPs enable full interoperability and backward and forward compatibility. They also provide the advanced telemetry and diagnostics, critical to maintaining an extremely resilient and serviceable network. We've been shipping the industry's highest-speed 800-gig PAM4 DSPs in high volume for several quarters and have begun sampling our next-generation 1.6T platform. We're seeing demand for connectivity between regional data centers accelerate as inference is deployed across multiple locations. As Ashish told you, Marvell has been a key enabler of this application with our DCI products, and we just announced our plan to demonstrate the industry's first 800ZR modules in October based on our new Orion coherent DSP. Looking at the future of optical connectivity, we are uniquely positioned in the industry with a leadership position in both PAM and coherent technology. We are also excited about the opportunity for our next generation of Ethernet switches, our 51.2T Teralynx 10 platform, which we announced earlier this year. We have begun sampling this product and we are seeing strong interest from customers. Last quarter, we told you how cloud customers are enhancing their AI offerings by building custom accelerators of their own. Trend is leading to a larger and faster-growing opportunity for Marvell's custom compute portfolio. We have won a number of custom silicon programs tied to AI and these are well underway to start ramping into volume production next year. Let me now talk about what we're seeing in storage in data center. As we expected, from a low base in the first quarter, we saw sequential storage data center revenue growth in the second quarter, and we are expecting modest sequential growth in the third quarter. However, storage end market demand remains significantly depressed and customer inventory remains high. As a result, the industry's expectations for a data center storage recovery have pushed out meaningfully. Looking ahead to the third quarter, we expect sequential revenue growth from overall cloud to accelerate above last quarter's performance, driven by continued strong growth from cloud AI, as well as standard cloud infrastructure. Demand for our AI products continues to grow at an extraordinary rate and we are working very closely with our customers to meet the rapidly evolving needs. On the other hand, enterprise on-premise is expected to continue to trend down. As a result, we are projecting overall data center revenue in the third quarter to grow in the mid-teens sequentially on a percentage basis. Turning to our carrier infrastructure end market. Revenue for the second quarter was in line with our guidance at $276 million, declining 3% year-over-year and 5% sequentially. Sequential and year-over-year decline were driven entirely by the wired portion of our carrier end market, reflecting ongoing demand weakness and inventory digestion at wired customers. In contrast, our wireless revenue continued to grow in the second quarter, building upon the 25% sequential growth we saw in the first quarter, and we are expecting additional growth in the third quarter. As a result of significant share and content gains for Marvell products in conjunction with the 5G upgrade cycle, we have grown our wireless revenue significantly over a multiyear period. While the full conversion to 5G in the world's installed base of wireless infrastructure will take many years, a number of regions are completing their initial phase of 5G deployments and are taking a pause in a challenging macroeconomic environment before they upgrade the balance of their networks. As a result, following an extended period of strong growth, we are expecting a significant sequential reduction in our wireless revenue in the fourth quarter. However, we expect that once customer and operator inventories normalize and carrier CapEx returns to more healthy levels, we can resume growth in our overall carrier end market and start to realize additional share gains. These will come from 5-nanometer base station designs we have won, but which are not yet in production. In addition, we expect the launch of our next-generation 800-gig Orion coherent DSP platform will drive long-term growth from the wired optical transport market. Moving to our outlook for the third quarter, we expect revenue from our overall carrier end market to grow in the low-single-digit sequentially on a percentage basis driven by wireless. Turning to our enterprise networking end market. Revenue for the second quarter was $328 million, declining 4% year-over-year and 10% sequentially. As we have been signaling for the last few quarters, we continue to see inventory corrections impact customer demand in this end market. We expect this inventory re-normalization to take a few quarters to resolve as customer balance sheets get worked down over time. While we deal with these market dynamics in the near term, I would note that enterprise networking has been an important contributor to Marvell's successful transformation to a leader in data infrastructure. The Marvell team has driven an extended multi-year period of exceptional revenue growth, with enterprise networking revenue essentially doubling over the last few years. This was enabled by a significant share in content gains, a testament to the consistent investment we've made in refreshing our enterprise networking product portfolio. As Ashish told you, we continue to introduce new products such as the industry's first 5-nanometer multi-gig Ethernet PHY transceiver. Looking ahead to the third quarter of fiscal 2024, we project our enterprise networking revenue to decline in the low teens sequentially on a percentage basis due to the market dynamics outlined earlier. Turning to our automotive and industrial end market. Revenue in the second quarter was $110 million above guidance, growing 32% year-over-year and 23% sequentially. Year-over-year growth was led by our automotive business, which continued to benefit from the growing adoption of Ethernet in cars. We also closed on a number of new automotive Ethernet design wins with multiple top 10 automotive OEMs during the quarter. Looking to the third quarter of fiscal 2024, we project revenue from our auto and industrial end market to be flattish sequentially and to continue growing year-over-year in the 30% range. Moving on to our consumer end market, revenue for the second quarter was $168 million, growing 2% year-over-year and 18% sequentially. Revenue was below guidance as deliveries for an end-of-life program were rescheduled to the third quarter. As a result, we are forecasting consumer end-market revenue to grow sequentially in the low teens on a percentage basis in the third quarter. In summary, we delivered revenue and earnings above the midpoint of guidance for the fiscal second quarter. We are forecasting revenue growth to accelerate in the third quarter, accompanied by gross margin expansion. We intend to remain disciplined on operating expenses to help us deliver strong operating leverage. Looking ahead, while inventory digestion in some end markets is taking longer to resolve, demand from AI applications continues to strengthen and Marvell is well-positioned to benefit from that trend. Based on our latest demand outlook for our electro-optics products, we now expect revenue from AI to exit this year at over a $200 million quarterly revenue run rate or $800 million annualized. This is well above what we had outlined last quarter. Put this in perspective, this would put us at the run rate we had previously communicated for all of next year. Looking forward between the ongoing strength from electro-optics and the expected ramp of multiple custom compute programs, we are expecting continued outsized growth from AI. Our results and outlook continue to validate our strategy to focus on developing the most advanced silicon for data infrastructure. The diversification in our end markets is serving us well, with strong growth from AI and cloud carrying us through the softening macro environment. With that, I will turn the call over to Willem for more detail on our recent results and outlook.
Willem Meintjes, CFO
Thanks, Matt, and good afternoon, everyone. Let me start with a summary of our financial results for the second quarter of fiscal 2024. Revenue in the second quarter was $1.341 billion, exceeding the midpoint of our guidance, declining 12% year-over-year and growing 1% sequentially. Data center was our largest end market, driving 34% of total revenue. Enterprise networking was the next largest end market with 24% of total revenue, followed by carrier infrastructure at 21%, consumer at 13%, and auto/industrial at 8%. GAAP gross margin was 38.9%. Non-GAAP gross margin was 60.3%, growing 30 basis points sequentially, driven by cost improvements, partially offset by a weaker revenue mix. Looking ahead, we expect gross margins to continue to improve in the third quarter and then to increase significantly in the fourth quarter. As Matt told you, the recovery in storage continues to push out, which is negatively impacting our product mix. However, in the fourth quarter, we project a significant improvement in our overall product mix to lead to stronger gross margin. We expect this improvement will be driven by our continuing growth in data center, while wireless carrier and consumer revenue declines on a relative basis. In addition, the Marvell team continues to execute well on our efforts to reduce costs. As a result, we continue to target non-GAAP gross margin returning to the bottom end of our long-term model of 64% to 66% in the fourth quarter. Moving on to operating expenses. GAAP operating expenses were $727 million, including share-based compensation, amortization of acquired intangible assets, restructuring costs, and acquisition-related costs. Non-GAAP operating expenses were $448 million, $7 million below guidance. We are pleased to report that we accelerated our cost reduction plan we outlined last quarter. We remain on track to execute the remainder of our cost reduction plan by the end of this fiscal year, as we communicated last quarter. Moving on to the rest of the income statement. GAAP operating margin was negative 15.7%. Non-GAAP operating margin was 26.9%. For the second quarter, GAAP loss per diluted share was $0.24. Non-GAAP income per diluted share was $0.33, $0.01 above the midpoint of guidance. Now, turning to our cash flow and balance sheet. During the quarter, cash flow from operations was $113 million. Operating cash flow was negatively impacted by an increase in DSO as well as severance-related cash restructuring charges. Our DSO increased 13 days from the prior quarter, primarily due to worse linearity as we ramp shipments on orders that were received well within lead time. CapEx was $111 million, which included a large number of leading node tape-outs that we expect to drive our future growth. As a reminder, our CapEx can be lumpy in any given quarter. We expect CapEx on average to be approximately mid-single-digits of revenue on a percentage basis. Inventory at the end of the first quarter was $1.02 billion, decreasing by $10 million sequentially. We returned $52 million to shareholders through cash dividends. Our total debt was $4.15 billion. Our gross debt to EBITDA ratio was 2.04 times, and net debt to EBITDA ratio was 1.83 times. During the quarter, we paid down $500 million of our total debt. Looking ahead, we will opportunistically explore accessing the debt capital markets to refinance our upcoming debt maturities. As of the end of the second fiscal quarter, our cash and cash equivalents were $423 million. Turning to our guidance for the third quarter of fiscal 2024. We are forecasting revenue to be in the range of $1.4 billion, plus or minus 5%. We expect our GAAP gross margin will be in the range of 45.6% to 48%. We project our non-GAAP gross margin will be in the range of 60.3% to 61.3%. We project our GAAP operating expenses to be in the range of $666 million to $671 million. We anticipate our non-GAAP operating expenses will be in the range of $435 million to $440 million. We expect other income and expense, including interest on our debt, to be approximately $48 million. For the third quarter, we expect a non-GAAP tax rate of 6%. We expect our basic weighted-average shares outstanding to be 863 million and our diluted weighted-average shares outstanding to be 869 million. As a result, we anticipate GAAP loss per diluted share in a range of a loss of $0.02 to $0.12 per share. We expect non-GAAP income per diluted share in the range of $0.35 to $0.45. In summary, for the third quarter, we are guiding for solid sequential revenue growth, further expansion in non-GAAP gross margin, and additional reductions in non-GAAP OpEx, all of which positions Marvell for strong operating leverage and earnings growth. In addition, following the paydown of $500 million in debt in the second quarter, we resumed buybacks in the third quarter. Operator, please open the line and announce Q&A instructions. Thank you.
Operator, Operator
We will now start the question-and-answer session. Our first question today comes from Tore Svanberg with Stifel. Please go ahead.
Tore Svanberg, Analyst
Yes, thank you. Matt, could you just elaborate a little bit more on the storage business? I am just trying to understand all the dynamics there. So, it sounds like there is a delay in the recovery, obviously, because there's still some inventories out there. But I'm also trying to understand the growth in storage associated with AI. I know these are different architectures, right? But if you could give us any more color on the digestion of the inventory on the compute side, coupled with how storage could benefit more on the AI side?
Matt Murphy, Chairman and CEO
Thank you for the question, Tore. Regarding storage, a good way to gauge the situation is to consider the feedback from end customers, as well as insights from both hard drive and flash companies, especially in relation to how long it's taking to deplete their inventory. Although it is taking longer than we initially projected, we expect a recovery eventually, possibly in the early part of next year or the first half, but we will have to see how that develops. On the AI topic, it's quite engaging. We are working on modeling this aspect, but we currently lack a strong framework for storage. We do have a clear perspective on how we fit into the overall landscape with our optical products, custom silicon, and networking. Overall, we believe the impact of AI on storage is net positive, though we do not have a particularly useful model to quantify it. AI should generally provide a supportive boost for storage, but this is still intertwined with the inventory issues being addressed at the cloud service providers.
Tore Svanberg, Analyst
Sounds good. Thank you.
Operator, Operator
The next question is from Timothy Arcuri with UBS. Please go ahead.
Timothy Arcuri, Analyst
Thanks a lot. Matt, AI revenue was initially projected at $400 million this year, but it's actually going to reach $200 million per quarter, meaning it will exceed $400 million for the year. Can you discuss how much of the final figure for the year, likely around $500 million to $550 million, will come from custom ASIC? Additionally, you mentioned that next year's AI revenue would be more than $800 million, and it will likely be significantly higher than that. Can you provide an updated estimate and clarify how much will be from custom ASIC? I'm estimating around $150 million, but I'd appreciate more details. Thanks.
Matt Murphy, Chairman and CEO
Thank you, Tim. I'm very encouraged by the upward demand trajectory in our AI segment. We expect to exit the year at a $200 million run rate, primarily driven by our electro-optics platform, DSPs, TIAs, and 800-gig drivers. There may be some contribution from custom silicon in the fourth quarter, but that will significantly ramp up next year. This is a positive indication because a quarter ago, when we made our estimates, we anticipated this growth. Exiting the year at next year's run rate is encouraging. This year, the growth is mainly from electro-optics, which will continue to grow next year, and we expect a meaningful increase in custom silicon as well. Currently, demand is rising so quickly that it's challenging to provide an exact number for next year, but it will clearly be larger than our previous estimates.
Timothy Arcuri, Analyst
I see there's a significant increase expected in the latter half of the year. Regarding the AI segment, will it maintain a steady growth rate throughout the year, or will it eventually level off as we move into next year? I'm just trying to clarify that. Thank you.
Matt Murphy, Chairman and CEO
I believe that in Q4, we're already at $200 million, and it will be at a higher run rate for each quarter throughout fiscal '25. This growth will continue on the optics side, and we will also add custom silicon. We’re definitely not suggesting that it will level off at that point; it’s coming much quicker, demand has improved significantly, and our supply chain team at Marvell is effectively sourcing the necessary components and capabilities. Overall, this creates an excellent foundation for next year concerning AI revenue.
Timothy Arcuri, Analyst
Great, Matt. Thank you so much.
Operator, Operator
Your next question is from Blayne Curtis with Barclays. Please go ahead.
Blayne Curtis, Analyst
Hey, guys, thanks for letting me ask the question. I wanted to ask, it's either Matt or Willem, I guess, on the gross margin. Can you just remind us what the headwinds are that you expect to resolve in Q4? Because I guess storage is taking longer, but I think the gross margin in electro-optics are quite good, so you would think that would be a tailwind. So, could you just walk us through the puts and takes on the gross margin guide, that would be helpful.
Willem Meintjes, CFO
Yeah, thanks, Blayne. Yes, certainly the optics is a tailwind. And then in addition to that, we've signaled that wireless carrier is really stepping down in the fourth quarter and consumer is also really stepping down. So, the combination of those things. And then in addition, our cost structure for our products is really improved. And the overhead on manufacturing and we see that benefit, it's really starting to flow through more significantly in the fourth quarter. So, storage certainly is a headwind for us as that recovery is pushed out.
Blayne Curtis, Analyst
Okay. Thank you.
Operator, Operator
The next question is from Matt Ramsay with TD Cowen. Please go ahead.
Matt Ramsay, Analyst
Thank you very much, everyone. Good afternoon. Matt, I believe you addressed the AI revenue discussion. I wanted to ask about the goals for custom silicon. We've talked about it previously in terms of $400 million and later $800 million, but that timeline has been extended. None of us have witnessed such significant shifts in capital expenditures related to Generative AI and spending behaviors from many of your large customers. We've noticed several custom compute ASIC projects in progress, some of which are aimed at CPU offload, whether through DPUs or SmartNICs, and others focused on custom silicon for smaller model inference at hyperscale. Could you share insights on where your engagements stand? What are you observing regarding timing and scale? The current capital expenditure environment has been quite volatile, especially concerning AI compared to traditional computing. Any updates on the custom compute programs and their timelines would be greatly appreciated. Thank you.
Matt Murphy, Chairman and CEO
Thank you for the question, Matt. Regarding the custom silicon opportunity, it fundamentally aims to deliver accelerated computing in a tailored format for various companies. The overall shift in capital expenditures towards accelerated computing is likely to benefit participants in that sector, including us. Some of this revenue overlaps with the AI figures I mentioned earlier, but the outlook appears positive. Last quarter, we discussed two key products linked to AI; one is in the sample stage and looks promising, expected to enter production soon. The second product has tape-out as scheduled in the second quarter and is also set to ramp up next year. The size and timing of these developments are still being finalized in collaboration with our customers. The overall trend in CapEx spending is a supportive factor for this business. Currently, we are tracking towards what we previously updated you on, and next year will depend on the extent of the transition and the volume of business generated. We are not making specific forecasts, but it is clear that our custom initiatives will be primarily driven by AI next year.
Matt Ramsay, Analyst
Got it. No, thank you, Matt. Appreciate it.
Operator, Operator
The next question is from Harsh Kumar with Piper Sandler. Please go ahead.
Harsh Kumar, Analyst
Yeah, hey. I think when you talked about gross margin drivers for fourth quarter, Matt, you mentioned that you might be expecting a significant recovery in revenues as well. Did I, a, understand that correctly? And if so, could you talk about what might be happening in the fourth quarter to drive that recovery?
Matt Murphy, Chairman and CEO
Yeah. You're saying overall, Harsh?
Harsh Kumar, Analyst
Yes.
Matt Murphy, Chairman and CEO
Your question is about our outlook. We're providing guidance one quarter at a time, but to give you some perspective, we've shared a lot of factors, including our commentary on carrier and wireless, particularly about the 5G run taking a pause. Overall, we expect revenues to increase again in the fourth quarter, aligning with what we previously mentioned regarding accelerated growth in Q3 and Q4. You can think of the fourth quarter as a continuation of the growth we've seen in Q3, driven primarily by cloud and AI, with AI being a significant factor. Additionally, as noted in our prepared remarks, standard cloud infrastructure revenues are also seeing solid growth throughout the year, not just from Q1 to Q2 and Q2 to Q3, but continuing into the fourth quarter and beyond. If you look at the numbers, you'll see that the data center is becoming a larger part of our revenue and AI is also growing more significant by the fourth quarter. Overall, Marvell, Inc. is experiencing an upward trend. A long time ago, we aimed to develop a diverse strategy in data infrastructure, targeting multiple end markets with a range of products including enterprise, 5G, automotive, and data center. Despite some quarters performing better than others in this dynamic environment, we are fortunate to have markets that have picked up at the right time. Even in a challenging macroeconomic climate, Marvell has continued to grow throughout the year, starting from our first quarter, and we expect that growth to continue through Q3 and into Q4, with shifts in the market mix. I know you'll develop your model, and we've provided enough information to piece it together.
Harsh Kumar, Analyst
No, we appreciate it, Matt. Thank you so much.
Operator, Operator
The next question is from Christopher Rolland with Susquehanna. Please go ahead.
Christopher Rolland, Analyst
Hey, thanks for the question. And this one is for Matt. Just kind of an amazing revision here on the electro-optics portion. Since you've seen this inflection, you probably done some more research here, how are you thinking about the attach rate for these products per GPU, call it, is it one for one? Are you thinking it could be two for one? I've seen some research that suggests depending on how many layers there are, could even be three per one. And is the revenue that we're talking about here, all 800-gig PAM4 DSPs? Is there anything else related in that as well? Thanks.
Matt Murphy, Chairman and CEO
Yeah. Thanks, Chris. Yeah, I think it's fairly similar to our view last quarter. We were pretty clear about the direct attach, which was the one-to-one you mentioned. Understanding that as you get to the upper layers of the network, there is more, and there is a range and you probably size the range. We're still I think refining our exact models there, so I don't know if I have an exact number to give you. But I think you're thinking about it the right way, starting at the direct attach and then building higher. And the second part of your question, it's all 800-gig for AI, at this point. We have strong traction on our next-generation products at 1.6T, which would be starting sometime next year. The products that are at frequencies lower than 800 gigabits are typically served for the traditional cloud infrastructure, which is also seeing a recovery as well. So, hopefully that's helpful.
Christopher Rolland, Analyst
Fantastic. Thanks, Matt.
Operator, Operator
The next question is from Ross Seymore with Deutsche Bank. Please go ahead.
Ross Seymore, Analyst
Hi, guys. I don't know if this one is for Matt or Willem. But I just wanted to talk about the puts and takes to next year's gross margin, and I know you're not guiding that far out with any specifics. But it seems like you're going to have a number of custom products that will be going up in the data center side, the storage side should recover at some point in time. But the general question is if custom products tend to be lower gross margin and storage and some other areas like enterprise eventually come back cyclically, how do we think about the puts and takes in your gross margin versus that 64% to 66% historical target range?
Willem Meintjes, CFO
Yeah, Ross, maybe I can start and Matt can add. So, when you look back, certainly over the last couple of years, we've grown both our carrier and ASIC business sort of faster than the rest of the business. And we were able to maintain our gross margin within our target range. We're targeting to get back to that 64% exiting this year and then to maintain that through next year. But clearly, it's sort of early to decide exactly how big the ASIC ramp is next year. Now if we do show outsized growth there, that would negatively impact our gross margin, but certainly our view is that, that would be very accretive to operating income and to EPS. But it's too early right now to know exactly the extent of that. Hopefully, that's helpful, Ross.
Matt Murphy, Chairman and CEO
And Ross, it's Matt. I want to add that the custom business has a lower gross margin, which we have been transparent about. Looking ahead to next year, it's a bit premature to make predictions since we are unsure about the recovery rate for storage, which is a significant factor. Additionally, within AI and cloud, we need to consider the revenue from optics compared to the custom sector. Automotive continues to perform well and grow, representing a higher-than-average gross margin category for us. We have various businesses with potential that may rise or fall at different times. As Willem mentioned, we consistently managed our gross margins within our targeted range, despite ongoing mix challenges. Recently, in Q1, Q2, and Q3, we faced a longer period of unfavorable gross margins compared to our traditional mix. However, the environment is dynamic, making it hard to predict when those markets will deplete their inventories and recover. At a high level, we believe we have a well-balanced portfolio of products and technologies with diverse business models that contribute to various gross and operating margin profiles. For instance, as Willem pointed out, even though some custom programs may yield lower gross margins than our corporate average, if they scale significantly, they can greatly enhance our operating margin and income. Ultimately, our focus is on increasing earnings per share and operating profits. So, this is a high-level overview, but we need a bit more time to clarify what next year will look like.
Ross Seymore, Analyst
Fair enough. Thank you.
Operator, Operator
The next question is from Vivek Arya with Bank of America. Please go ahead.
Vivek Arya, Analyst
Thank you for taking my question. Matt, I was hoping you could help us size how big storage is. Currently, in the quarter you reported both kind of data center and outside of data center. And I forgot whether you mentioned whether that's going to be up/down/flat in Q3 and Q4, or can it kind of hold at these low levels? And what are you looking for to inform you as to when it starts to grow sequentially, like how much excess inventory is out there, or do you think that it can actually hold at these low levels? So, just help us kind of set what the right baseline view is of storage as you get into next year.
Matt Murphy, Chairman and CEO
Thank you, Vivek, for the great question. It's quite significant. Let me update you on our current position. In Q1, our storage revenues were very low, particularly in the data center storage sector. The consumer segment has performed relatively stable thanks to some specific applications. However, the primary driver for change is the data center side. In Q2, we saw a recovery, starting from a low point of around $100 million. We mentioned in our prepared remarks that Q3 would show a modest recovery. Although this growth is positive, it’s still not aligning with our earlier expectations regarding the year-end exit rate. Most feedback from customers indicates that a recovery will be more pronounced in the first half of '24. The encouraging news is that we've moved away from the low point, with growth in Q2 and anticipated growth in Q3 and Q4. However, the growth rate isn't quite as steep as we envisioned. Upon consulting with our customer base, it’s challenging to pinpoint the exact status of inventory and timing. Nonetheless, we are aware of significant activity in qualifying next-generation drives and technologies, with a notable total cost of ownership benefit. We haven't noticed any fundamental changes, although we continue to analyze the situation. Ultimately, we believe the volume of exabytes being shipped or consumed will return to previous levels and continue to grow, driven by trends that have persisted for the last two decades. AI may provide additional support in this area. However, given the complexities of the supply chain, it's difficult to fully understand current dynamics at the customer level and the timing of recovery. We will continue to pay close attention to customer feedback and adjust accordingly. I hope this information is useful.
Vivek Arya, Analyst
Yes. Thank you, Matt.
Operator, Operator
The next question is from Srini Pajjuri with Raymond James. Please go ahead.
Srini Pajjuri, Analyst
Thank you. Hi, Matt. My question is about the custom silicon side. You mentioned two AI programs that you expect to ramp up next year. Can you discuss the design win pipeline? Also, I understand it’s challenging to forecast revenue opportunities this early, but could you provide some insight? We have heard that some of these programs can get canceled, so I want to ensure you have visibility that these programs are legitimate for next year. Additionally, if you could touch on other opportunities beyond those two programs, that would be helpful. Thank you.
Matt Murphy, Chairman and CEO
That's a great question. To address the reality of the situation, there were significant shifts in the macroeconomy that impacted cloud companies about two quarters ago, prompting them to take actions on headcount and expenses and reassess their program priorities. In our calls in Q1 and Q2, we updated investors on how those changes were unfolding. We observed a transition from traditional cloud infrastructure towards AI, which has clearly emerged in the first half of the year. These AI programs not only have funding but also come with a strong sense of urgency. We're now discussing production timelines and ensuring that our capacity aligns appropriately. This gives us confidence about our progress, and as we move forward, we expect to provide more visibility in the coming quarters. We're optimistic about our programs and have a clear understanding of our customers' goals. Regarding the overall opportunity size, Marvell's total design opportunity pipeline is heavily focused on the data center segment, which accounts for a large part of our open funnel. Specifically, the AI segment continues to grow significantly. Our opportunities stem from various technologies, including our PAM products and AEC offerings, as well as custom compute silicon programs, which are in high demand. There's also strong interest in our Ethernet switching platform, and we have exciting new technologies like silicon photonics for the data center. Overall, the conversations we're having with customers demonstrate a desire to maximize the benefits of our solutions, focusing on total cost of ownership, power, performance, and optimizing their architecture. As we engage with our major cloud customers and partners, we're presenting ourselves as a unified company with a comprehensive suite of solutions capable of processing, moving, storing, and securing data in the cloud. This collective approach is very promising, and there are numerous opportunities ahead of us.
Srini Pajjuri, Analyst
Thanks, Matt.
Operator, Operator
The next question is from Harlan Sur with JPMorgan. Please go ahead.
Harlan Sur, Analyst
Hi, good afternoon. Thank you for taking my question. There's been a lot of attention on your electro-optical business related to AI. However, I believe the larger part of your electro-optical business focuses on the ongoing upgrade cycles across the entire cloud data center sector. These trends still appear to be ahead for the team. There remains one more major cloud provider that has yet to upgrade to 400-gig in its data center. Additionally, there are the 800-gig 1.6T optical upgrades needed across the other three major providers. On the DCI front, you are still ramping up 400ZR DCI, but I believe there are still two more cloud customers in development. Are these sweeping upgrade cycles likely to align and start next year, possibly with some beginning in the latter half of this year? Maybe this connects to your comments regarding strength in the broader cloud market. I'm just trying to understand the ramp-up profile for these upgrade cycles.
Matt Murphy, Chairman and CEO
I think you summarized it well. The AI segment has clearly become significant, and we have an idea of the revenue expectations for the fourth quarter. It's impressive how quickly that's grown. Additionally, the standard cloud infrastructure aspect is recovering nicely from the inventory built last year, which is now becoming beneficial as it decreases. You mentioned that a major cloud company still needs to complete its upgrades, which is also ahead of us. As we transition each technology in terms of speed, there is always an increase in average selling price, which is reasonable because we're offering twice the bandwidth at not twice the price. So, we are on a positive path to share the benefits with our customers. Looking ahead to next year, while AI will lead, it will also be an important component of the overall infrastructure growth in the following year. Furthermore, the regional data center products, particularly the ZR products, are performing extremely well this year, with significant upside from Generative AI. We mentioned in our prepared remarks and the press release about 800ZR, which came out recently. There will be more updates on that. We're expecting another frequency enhancement to arrive sooner than we anticipated six or nine months ago, as the bandwidth between data centers is now becoming a limitation. So, there are substantial opportunities driving our cloud growth that extend beyond just AI, and we're observing a lot of this in the second half and expect it to continue into next year. We have multiple avenues for Marvell's overall growth.
Harlan Sur, Analyst
Perfect. Thank you, Matt.
Operator, Operator
The next question is from Ambrish Srivastava with BMO. Please go ahead.
Ambrish Srivastava, Analyst
Hi. Thank you very much. Willem, I had a question on the cash flow side. I was trying to see if there was a discernible pattern of seasonality, doesn't seem to be. But you gave a reason for DSOs and impacting the (CFO]. Could you kind of just walk us through how should we think about cash flow for the rest of the year?
Willem Meintjes, CFO
Yes, definitely. Thank you, Ambrish. This quarter, our days sales outstanding were somewhat affected by linearity. We anticipate a strong recovery in the third quarter along with some normalization. As we've highlighted, we are concentrating on reducing our days of inventory, which we have been successfully doing throughout this year, and we plan to keep this momentum going for the remainder of the year. So, you can expect significant improvement in the second half.
Ambrish Srivastava, Analyst
Okay. And my quick question for you, Matt. We're hearing a lot of companies are talking about the pushouts of CapEx, general-purpose CPU towards AI. I know you addressed it in some ways earlier. But are you seeing that manifest in your business that certain projects are being sacrificed at the AI alter?
Matt Murphy, Chairman and CEO
It's an interesting way to put it. Yes, I think there are two aspects to consider. One is whether there has been any project reprioritization within those companies, and the other is whether it is currently impacting our business. From our perspective at Marvell, I can't speak for everyone else, but we addressed this issue at the end of last year and early this year. There was significant reprioritization due to budget constraints as companies sought more efficiency. This was also aligned with the emergence of ChatGPT and the recognition of the huge potential for productivity gains through AI, which sparked a competitive rush. We feel positive about our work in progress and the decisions we made regarding reprioritization. Some of these decisions contributed to our lower custom silicon revenue this year, as we discussed a few quarters ago. These programs experienced delays, but we've moved beyond that, and now we’re fully engaged in the AI developments. Looking at our current business, in Q2, we saw a healthy increase in overall data center revenues, despite on-prem services being down. AI revenues saw a substantial rise, and we noted that standard cloud infrastructure revenues increased and are expected to increase again in the third quarter, continuing through the fourth quarter and into next year. We've managed to navigate these changes, likely because we focus more on networking and connectivity rather than just selling CPUs. The demand for network bandwidth is essential because many large data centers are multi-tenant, housing AI and specialty-built servers and storage. Everything needs to function cohesively with adequate bandwidth to support computing power. In terms of challenges, the storage aspect, which impacts our data center near-line business, has been a consistent issue. However, overall, we're seeing significant growth, which we anticipate will continue to be a growth driver for us next year.
Ambrish Srivastava, Analyst
Got it. Thank you, Matt.
Operator, Operator
The next question is from Chris Caso with Wolfe Research. Please go ahead.
Chris Caso, Analyst
Thank you. I’d like to delve deeper into the enterprise on-premise aspect. Can you provide more details on what that includes, especially since it appears to be declining? Does this also cover fiber channel? Do you believe this signals a bottom for that business? It may be challenging to determine when it will start to improve.
Matt Murphy, Chairman and CEO
Thank you for the question, Chris. The enterprise on-premise business has continued to decline. The reality is that overall enterprise server shipments have likely been down for about eight quarters. The end markets are not performing well, primarily driven by fiber channel, along with some Ethernet and NIC products. While I’m not saying we’ve reached the bottom, it has to be close since eventually, customers will need to purchase servers and address their inventory levels. This situation has been challenging, and we might need a little more time—perhaps another quarter—before things normalize. Despite this, the overall data center inventory adjustments and weaknesses in the end market are being overshadowed by the growth in AI and cloud infrastructure, which continues to show strong performance into Q3 and is expected to increase further in Q4, even without a significant recovery in the on-prem side. Looking at fiber channel specifically, it has been a stable business for us since the Cavium acquisition, and I see no reason why it wouldn’t recover to its previous levels. When that happens, it will be beneficial for us as well. I hope this clarifies the situation.
Chris Caso, Analyst
Thank you.
Operator, Operator
The next question is from Quinn Bolton with Needham & Company. Please go ahead.
Quinn Bolton, Analyst
Hey, Matt, I wanted to ask, you talked about the electro-optics driving most of that $200 million of AI revenue. Can you give us some sense how much of that is InfiniBand versus Ethernet within those 800-gig modules? And do you guys see any share difference between your position in Ethernet versus InfiniBand? I know your share is very high in general in that market, but just wondering if you would say your share is higher either in the Ethernet or InfiniBand. And then I got a follow-up.
Matt Murphy, Chairman and CEO
Yeah, maybe the simple answer is, first of all, as you point out, we're agnostic. We participate broadly in the connectivity independent of whether it's InfiniBand or Ethernet. And I'd say that we continue to have very high market share and penetration into both those applications. And the demand we're seeing is broad-based from both of those as well as multiple customers driving it. So, it's actually really kind of hard to break it down exactly. And I don't know that it's super helpful, because overall, the shares still very healthy, and we're agnostic. So, the same module can kind of be used in either one.
Quinn Bolton, Analyst
Got it. Okay. So it's pretty broad-based across both. The follow-up question is just you mentioned 51.2 terabit switch and starting to sample. Just as you look out to calendar '24, fiscal '25, do you see that starting to get deployed at the U.S. hyperscaler partners?
Matt Murphy, Chairman and CEO
Yes, I believe we will see some contribution from that product area next year. The larger ramp is a bit further out, but I’m pleased with the open opportunities in that technology. I’m very satisfied with the team. The architecture and key components initially came from Innovium, which we acquired in 2021. We integrated it into the Marvell 5-nanometer design flow using our SerDes, IP platform, and new product development process. This approach combines the best of both worlds, and the teams executed exceptionally well on the chip. Consequently, it’s instilling confidence in our customers that we can support them not only in terms of features, such as very low latency within the appropriate power envelope, but also in high-volume manufacturing, which is quite challenging for these advanced switches and products. We’ll see how things progress, but the pipeline looks strong. This will be a significant upgrade cycle for the industry, and I expect it to be a positive development for those involved in that ecosystem.
Quinn Bolton, Analyst
Thank you.
Operator, Operator
I think we'll take one more question, Gary, and then we can conclude. And that question will be from Gary Mobley with Wells Fargo Securities. Please go ahead.
Gary Mobley, Analyst
Hey, guys. Thanks for taking my question. I don't think there's been a mention so far the influence on an extra week in the fourth quarter. Matt, I want to verify your expectation of sequential revenue growth in the fourth quarter. Is that independent of the extra week or inclusive of the extra week? And Willem, the target of OpEx for the fourth quarter, I think you said $430 million with the extra week. Is that still the target?
Willem Meintjes, CFO
That's correct.
Matt Murphy, Chairman and CEO
Sure. Regarding revenue, based on my long experience, when we have a 14-week quarter, you typically don't see a significant increase in revenue, but you definitely need to account for the additional expenses. So, we aren't assuming that the extra week in Q4 will lead to better growth. That's just how the market behaves. Additionally, Willem provided the operational expense run rate, which, if you adjust for a 13-week period, comes to about $420 million.
Gary Mobley, Analyst
Got it. Thanks, again, guys.
Matt Murphy, Chairman and CEO
Yeah.
Operator, Operator
This concludes our question-and-answer session. Thank you for attending today's presentation. You may now disconnect.