Good afternoon, and welcome to Marvell Technology Fiscal Q1 2023 Earnings Conference Call. All participants will be in listen-only mode. Should you need assistance, please signal a conference specialist by pressing the star key followed by zero. 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 the conference over to Mr. Ashish Saran, Senior Vice President of Investor Relations. Please go ahead.
Thank you, and good afternoon, everyone. Welcome to Marvell's Q1 fiscal year 2023 earnings call. Joining me today are Matt Murphy, Marvell's President and CEO, and Jean Hu, our CFO. Let me remind everyone that certain comments made today may 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.
With that, I'll turn the call over to Matt for his comments on our performance. Matt?
Thanks, Ashish, and good afternoon, everyone. In the Q1 of fiscal 2023, the Marvell team drove another record level of revenue at $1.45 billion, exceeding the midpoint of guidance, growing 8% sequentially and 74% year-over-year. We saw continued strength in bookings in all our data infrastructure and markets. Higher revenue achievement was primarily driven by our data center market, with additional strength from carrier infrastructure and automotive results, both of which were also above forecast. Due to supply chain-related impacts, results from our enterprise networking market were below our guidance. However, growth was still very strong, with revenue growing a robust 64% year-over-year and 9% sequentially. The Marvell operations team did a great job in navigating a tight supply environment that was further compounded by COVID-related manufacturing challenges at our suppliers in certain geographies.
Our team's efforts were a key enabler of our Q1 revenue exceeding the midpoint of our forecast. We continue to make progress in securing additional capacity with our strategic partners to enable sustained revenue growth. Let me now move on to discussing our five end markets, starting with the data center. In our data center end market, revenue for the Q1 was $640.5 million, beating our guidance, having grown 12% sequentially and 131% year-over-year. Strong performance was broad-based, with multiple product lines contributing to excellent results. Cloud continues to be a source of Marvell's strength in the data center. We are enabling our hyperscale customers to add new use cases, bringing the benefits of AI and machine learning to businesses, increase process automation and productivity, and drive deeper relationships with their customers.
Let me now discuss a number of Marvell product cycles driving strong growth in cloud, starting with electro-optics . We are seeing strong demand for our PAM solutions inside data centers and ZR pluggables between data centers. Inside cloud data centers, leading tier one customers are currently driving volume deployment of our PAM-based 200 gig and 400 gig solutions. The rest of the market has plans for starting deployments later this year and next year. As bandwidth requirements continue to grow, we expect the role of PAM-based electro optics to expand, replacing legacy solutions and as a result, to grow our opportunity. We believe that the next generation of more powerful server CPUs will accelerate the need for PAM technology. In addition to this expansion, we are also increasing our content per module with our next generation of higher speed solutions that we are now starting to ship.
Driven by the growth in AI deployments, our Q1 results benefited from a ramp in volume shipments of our 800G PAM solutions at two large customers. The bandwidth expansion inside data centers is also driving a significant increase in connectivity between data centers, creating a growing opportunity for our 400 ZR pluggable optics. The team is achieving great success by driving the adoption of these solutions at multiple customers, and we are projecting strong revenue growth from these products. Moving on to compute, our cloud optimized design win momentum continued in the Q1 , and we won a custom SmartNIC at a hyperscale customer. We are seeing more adoption of DPU-based architectures inside data centers, a trend that we are ideally positioned to address with our OCTEON platform, which is now in its tenth generation.
This design win is one more example of the growing demand for cloud-optimized silicon, which we see as the largest incremental growth opportunity for Marvell inside data centers. We are confident that we are uniquely positioned to win these opportunities with our leading portfolio of compute, networking, security, storage, and high-speed electro-optics IP delivered on our 5-nanometer platform. Moving on to storage within the cloud. In the Q1 , cloud demand for high-capacity storage continued to increase, driving solid growth for our nearline HDD controllers and preamplifiers. Our SSD controllers also contributed to strong year-over-year revenue growth. We are now reaping the benefits from our data center SSD business, which we built from the ground up starting in 2016.
We increased investment in critical IP development, accelerated our process technology cadence, and quadrupled our firmware team. Working closely with the leading cloud companies, we optimized our SSD controllers to address their quality of service and security requirements. We developed state-of-the-art error correction for the most advanced NAND flash technologies, coupled them with our in-house low-power PHYs, and accelerated our PCIe roadmap. While the leading SSD devices today are shipping PCIe Gen 4 solutions, Marvell has already won Gen 5 data center sockets at three key NAND OEMs. Additionally, we started investing in PCIe Gen 6 in 2019. This development, coupled with our 5-nanometer technology platform, has enabled us to win a key NAND OEM for their Gen 6 SSDs.
Our data center storage business has been built on a long and sustained period of technology investment, deep system knowledge, extensive customer relationships, and very sticky custom firmware delivered on our advanced process technology platform. We expect our position to continue to strengthen based on our proven technology platform and multiple secured design wins in the next generations of data center SSDs. We anticipate that the next big evolution in cloud data centers will be the adoption of CXL, or Compute Express Link, an industry standard for connecting processors, accelerators, and memory, and we are planning to enable that trend. Last week, we held a tech talk in which we explained the fundamentals of CXL technology and its role in bringing new levels of performance to next generation cloud data centers.
If you were unable to attend, I would encourage you to watch the recording posted in the investor relations section of our website. We described how silicon components based on CXL will facilitate new cloud architectures by addressing the multiple memory scaling challenges in current data centers. Put simply, CXL will finally allow DRAM memory to escape the constraints of being tied down to a single compute device, such as a CPU, and become a shared and pooled resource, utilizing the well-established PCIe fabric as the interconnect. Marvell is uniquely positioned to address the CXL opportunity given our deep relationships with memory OEMs, our growing position within hyperscale customers, and our advanced PCIe roadmap. We recently bolstered our efforts in this area with the acquisition of Tanzanite, a leading developer of advanced CXL technologies.
Their complementary IP and world-class team adds more resources to Marvell, accelerating our CXL roadmap to address opportunities in closed design wins, which are right in front of us at multiple customers. We see opportunities for a host of new products, including CXL expanders, pooling devices, switches, and accelerators. In addition, we see the potential to embed CXL IP in a broad range of our data center products, including ASICs, custom compute engines, DPUs, electro-optics, retimers, SmartNICs, and SSD controllers. We see a multi-billion-dollar SAM expansion opportunity driven by CXL over time, and I look forward to updating you on our progress. Moving on to our expectations for the Q2 of fiscal 2023 from our data center end market, we are projecting continued growth to layer on top of the strong Q1 results.
We project our data center revenue in the Q2 to grow sequentially in the low single digits % and year-over-year grow approximately 50%. We expect revenue from cloud to grow significantly faster than the on-premise market, both sequentially and year-over-year in the Q . As we progress into the second half of this fiscal year and beyond, we are looking forward to additional incremental contributions from our cloud Ethernet switches and to ramping our large set of cloud-optimized custom design wins. Marvell's unique ability to offer all the critical data center IP under one roof, designed to work seamlessly together in a custom solution is proving very attractive to customers. As a result, our engagements with hyperscalers have moved well beyond pure ASIC programs to include various combinations of Marvell IP delivered in solutions tailored to each cloud's unique requirements.
Turning to our carrier infrastructure end market, revenue for Q1 was $252 million, above our forecast, growing 5% sequentially and 50% year-over-year. We previously reported a substantial sequential step-up of over 30% in our 5G business in the fourth quarter of fiscal 2022. From this strong base, it was great to see both sequential and year-over-year revenue growth for carrier continue in the first quarter. The growth in 5G deployment, combined with Marvell product ramps at multiple base station customers, continued to fuel strong growth in this end market. In wired, we are seeing strong demand for our 400 gig coherent electro-optics portfolio, driven by rapid adoption in the metro and long-haul carrier markets. Coherent technology is critical to enabling high-speed data transmission across long distances to meet the ever-increasing demand for bandwidth from operators.
Earlier this week, we announced that we have shipped over 100,000 400G coherent DSPs, positioning us as the leading merchant provider of these products. We are also aggressively focused on developing and launching our next generation of coherent products. This is a very key piece from the Inphi acquisition, as the same technology powers the 400 ZR DCI cloud market you heard about earlier. Looking ahead to Q2 , we expect revenue from the overall carrier end market to grow in the high single digits sequentially on a percentage basis, while year-over-year growth is expected to remain strong at approximately 40%. Moving on to our enterprise networking end market. Revenue for Q1 was $286.6 million, growing 9% sequentially and 64% year-over-year, as demand remains strong in this end market.
While this end market represents our highest delinquency relative to the size of the business, we remain strongly focused on improving our ability to supply more products to our enterprise networking customers to help them meet their growing demand. Our growth has been driven by share gains and our increase in content, starting to materialize as our customers began shipping their new platforms to address enterprise network modernization. We have seen a large increase in the adoption of our multi-gigabit PHYs, which have a significantly higher selling price compared to our gigabit products. We expect the penetration of multi-gigabit ports will continue to increase, a tailwind to our business. Our strong growth in enterprise networking is primarily the result of our own unique product cycles.
Looking beyond this quarter, we expect growth to continue, leveraging our refresh switch and PHY portfolio and incremental revenue from the ramp of custom silicon and our OCTEON ARM-based DPUs displacing alternative architectures. In Q2 of fiscal 2023, we expect a continuation of strong demand for our products from the enterprise networking end market and further improvements in supply. As a result, we are projecting revenue to be up sequentially in the mid-teens % and year-over-year growth of approximately 45%. Turning to our automotive and industrial end market. Revenue for Q1 was $89.3 million, growing 12% sequentially and 94% year-over-year. All of the sequential growth came from our automotive business, which drove over 50% of the total revenue from this end market.
Although we are still unable to fully satisfy the growing demand for our Brightlane Auto Ethernet solutions, the results, supported by incrementally better supply solutions, exceeded our guidance. Our auto revenue growth is being driven primarily by our new product cycles. The adoption of Marvell Ethernet technology in cars is continuing to increase. As OEMs design in higher speed solutions to address the increase in bandwidth, our dollar content per car is continuing to grow. We are also winning new customers and growing our content at existing customers. We now have Ethernet design wins at 8 of the 10 largest auto OEMs worldwide and 36 OEMs in total.
Looking ahead to Q2 of fiscal 2023, for the combined auto and industrial end market, we are projecting sequential revenue growth in the mid-single digits on a percentage basis, while year-over-year growth is expected to be over 60%. We expect strong growth to continue from our automotive business, with revenue projected to more than double year over year. Moving on to our consumer end market. Revenue for the first quarter was $178.5 million, growing 7% year over year. The growth in this end market is being driven by our SSD controllers, partially offset by declines in our PC HDD business. On a sequential basis, revenue declined by 4% in our consumer end market, below our forecast for a flattish outlook due to a reduction in demand from the PC HDD market.
While we are not the bellwether of global PC demand, as it represents a relatively small amount of our revenue, we did see a rapid change in tone from the PC market, and we expect the weakness to continue. From our perspective, the shift from HDDs to SSDs in the notebook PC market is almost complete, and our revenue from notebook HDDs in Q1 of fiscal 2023 was less than 1% of consolidated Marvell revenue. Looking ahead to Q2 of fiscal 2023 in our consumer end market, we expect revenue to sequentially decline in the mid-single digits% on a percentage basis and be flattish year-over-year. In closing, we delivered record results for the first quarter and are guiding for continued strong growth in Q2 .
We are seeing robust demand for our products, and our design win momentum remains strong. With 88% of our revenue coming from data infrastructure, we are confident that Marvell's favorable end market exposure makes us one of the best-positioned semiconductor companies to benefit from strong secular growth trends. In Q2 , at the midpoint of the range, we are guiding our revenue to grow by 41% year over year, which is almost entirely an organic comparison. We expect the strong operating leverage in our business model to drive non-GAAP EPS at the midpoint of guidance to grow by 65% year over year, significantly faster than our projected growth in revenue. As you heard throughout this call, our unique product cycles have been a big part of our above-market revenue growth.
As we look into the second half of this year and beyond, we are confident that our unique growth drivers in our cloud, 5G, auto, and enterprise networking end markets and our strong track record of execution through economic cycles will continue to be a source of strength. You'll remember from our prior calls and our Investor Day presentation that we have already won a significant number of design wins, which we are projecting will add a substantial amount of incremental revenue to Marvell going forward. While the external environment is challenging, Marvell's business continues to be strong, including strong bookings in our core data infrastructure end markets. As we move forward, we intend to continue to act as we always have, be diligent about changes, react promptly, and manage our business and costs aggressively.
On behalf of Marvell's leadership team, I thank our employees for their dedication in driving stellar results. Our employees are the backbone of the company, and we are excited that Marvell recently ranked third overall on the list of best places to work in the Bay Area by the San Francisco Business Times and Silicon Valley Business Journal. Receiving this award is so special because it is the employees who decide the winners. We have an outstanding team, and this award reflects our culture, values, and dedication to creating a collaborative workplace which fosters creativity and innovation. With that, I'll turn the call over to Jean for more detail on our recent results and outlook.
Thanks, Matt, and good afternoon, everyone. I'll start with a review of our financial results for Q1 and then provide our current outlook for Q2 of fiscal 2023. Revenue in Q1 was $1.447 billion, exceeding the middle point of our guidance, growing 8% sequentially and 74% year-over-year. The data center was our largest end market, driving 44% of consolidated revenue. Enterprise networking was the next largest with 20% of total revenue, followed by carrier infrastructure at 18%, consumer at 12%, and auto industrial at 6%. GAAP gross margin was 51.9%. Non-GAAP gross profit was $947 million, or 65.5% of revenue. Another record driven by rich product mix.
GAAP operating expenses was $681 million, and include the cost of share-based compensation expenses, amortization of acquired intangible assets, legal settlements, and acquisition and divestiture-related costs. Non-GAAP operating expenses were $435 million, reflecting the increase in R&D headcount and project expenses to execute design wins across all our data infrastructure and the market. As we outlined at our Investor Day last year, we plan to continue to invest in R&D given the tremendous opportunities we have in front of us, while keeping OpEx growth well below top-line revenue growth to drive to our long-term target model. GAAP operating income was $70 million. We achieved record non-GAAP operating profit over $512 million, up 123% from a year ago. Our non-GAAP operating margin was 35.4%.
During Q1 , our GAAP income tax expense was $205 million. This included the impact from a one-time non-cash tax charge over $214 million, offset by other tax benefits in the quarter. The $214 million non-cash tax charge was due to a remeasurement of our net deferred tax asset as a result of obtaining an extension of a lower foreign jurisdiction tax incentive. For the first quarter, GAAP loss per diluted share was $0.20. Non-GAAP income per diluted share was $0.52, up 79% year-over-year, exceeding the middle point of our guidance. Now turning to our balance sheet and the cash flow. During the quarter, cash flow from operations was $195 million. We increased our working capital investment to support the top line revenue growth in a very tight supply chain environment.
The increase in working capital was primarily driven by increase in inventory, especially raw materials and work in progress, to help us manage through ongoing supply chain disruption. We also made $86 million in payments during the quarter to secure long-term capacity with our suppliers. As of the end of the first fiscal quarter, our cash and cash equivalents were $465 million, and our long-term debt was $4.5 billion. Our gross debt to EBITDA ratio was 2.3 times, and our net debt to EBITDA ratio was 2 times. We returned $66 million to shareholders through $51 million in cash dividends and $15 million through share repurchases during the first quarter.
Subsequent to the end of the first quarter of fiscal 2023 or the first three weeks of Q2 , we purchased an additional $50 million of our shares through our 10b5-1 program. We are pleased that we have restarted the share repurchase program as we believe it's in the best interest of the company and our long-term shareholders. In summary, the Marvell team executed exceptionally well, delivering accelerated top line revenue growth and strong earnings expansion significantly faster than revenue growth. Turning to our guidance for Q2 of fiscal 2023, we expect the following results. Revenue in the range of $1.515 billion ±3%. GAAP gross margin in the range of 49.6%-51.9%. Non-GAAP gross margin in the range of 65%-65.5%.
GAAP operating expenses to be approximately $669 million. Non-GAAP operating expenses to be approximately $435 million, which assumes that we will complete the acquisition of Tanzanite in Q2 . For Q2 , we expect a Non-GAAP tax rate of 6%. We expect our basic weighted average share outstanding to be 853 million, and our diluted weighted average shares outstanding to be 862 million.
As a result, we expect GAAP earnings per share in the range of $0.02-$0.10. We expect non-GAAP income per diluted share in the range of $0.53-$0.59. Operator, please open the line and announce Q&A instructions. Thank you.
We will now begin the question-and-answer session. To ask a question, you may press star, then 1 on your touchtone phone. If you are using a speakerphone, please pick up your handset before pressing the keys. To withdraw your question, please press star then 2. In the interest of time, please restrict yourself to one question only. If you have additional questions, please rejoin the queue. Our first question is from Toshiya Hari with Goldman Sachs. Please go ahead.
Good afternoon. Thank you so much for taking the question and congrats on the strong execution. Matt, I was hoping you could speak to your expectations for the second half, both from a demand perspective and a supply perspective. I think last quarter there was a little bit of confusion as it pertains to the full year outlook. If you can kind of confirm how you're thinking about your overall business, but particularly, data center into the second half, that would be super helpful. If you have the supply to support the strong outlook? Thank you.
Thanks, Toshiya, and great to hear from you. So we're pleased with the results and execution in Q1. As you saw, we had a very strong quarter in data center, which I'll come back to in a minute. Just to kind of reframe from the demand side in all of our end markets except consumer, we see very strong demand trends, Toshiya. Bookings have been very strong in the last two quarters. Demand, our demand outlook for the year continues to be strong. With respect to data center specifically, we did have a strong Q1, better than expected.
We still have pent-up demand and lack of supply there for Q2 , but we expect actually a re-acceleration in the data center in terms of our growth in Q3 and Q4, i.e., the second half. On top of that we have incremental new design wins that we've secured over the last 12 to 18 months, which we've sized roughly at, call it $400 million incremental next year and $800 the year after. That's actually gonna start layering in as well in the second half. From our perspective, ex-consumer demand continues to be very strong. The data center in particular, is gonna have a great first half and a great second half.
Relative to supply, we really, first of all, credit our operations team for doing a great job in the Q1 , navigating all of the different impacts going on, in particular in China, and delivering a strong quarter. Looking forward, we're starting to see the benefit of all the investment we put in over the last year into our supply chain, whether that be the improvement in systems and tools, the relationships and, the capacity that we've secured. We believe that that's going to continue, and we do see a much improved supply outlook, not only in the short term, as you can see from our results, both beating in our Q1 and then guiding strongly for Q2, but also in the second half, Toshiya.
Again, I think that's the result of a lot of hard work and transformation in our supply chain to meet our growth plan. In summary, we have very strong demand, driven by data center, but also in 5G and automotive, and the other things we've talked about. We've seen real improvements in our visibility into the supply we're gonna get to go meet that customer demand.
Great. Thank you for all the details.
Yep.
The next question is from Chris Caso with Raymond James. Please go ahead.
Yes, thank you. Wondering if you could just talk a little bit about the enterprise networking business. I guess it came in a little below your expectations, but yet you're still speaking about supply constraints in that area. Is that a situation where, your customers are supply constrained elsewhere and you're waiting for them? Maybe just give some explanation of the dynamics in that part of the business.
Thanks, Chris. This was definitely a quarter where the mix wasn't exactly where we thought it was gonna be from when we guided. As you can see, we overachieved in data center and we underachieved on the enterprise. As I pointed out in my comments, the delinquency and kind of the unfulfilled backlog relative to the size of the business is actually still the highest in enterprise. This was really a case of just us being unable to execute our supply plan for the quarter, nothing to do with demand. In fact, the demand from our enterprise customers continues to be very robust looking into the Q2 and throughout the year. It was nothing to do with them. They want more of our product.
You can see somewhat of a catch-up in Q2 with a very strong guide, sequentially into the Q2 . We see that continuing, Chris. I know there's been a lot of noise out there around Enterprise, but from our perspective in talking to our customers, demand remains very strong. In fact, we're behind in terms of our supply to catch up. I think consistent with what Toshiya was asking, we are looking to improve that in the Q2 and also through the second half.
Thank you.
Yep.
The next question is from Gary Mobley with Wells Fargo Securities. Please go ahead.
Hey, everybody. Thanks for taking my question. Congratulations on the good results. Wanted to ask about the supply situation and you just articulated improving supply coming online for the balance of the year. To what extent is that being driven by you perhaps giving your foundry partners a better forecast? How much of it is being driven by maybe you throwing some more money their way in prepayments? And to what extent is it being influenced by some shifting of fungible capacity, the semiconductor market is no longer universally strong, so maybe you're benefiting from redistribution of supply?
Excellent question, and I think there's probably a piece of all of that. I think when I look back over what we've accomplished, one is we had to do a fairly significant education process across the supply chain of Marvell's prospects, how all the different pieces have come together from the different M&A we've done, our end market focus, et cetera. I think we did a good job of that, of getting the broader supply chain ecosystem to understand the Marvell opportunity. To a large extent, they're betting on us, and they're supporting our growth. You can see that, right, just in terms of the year-over-year growth we're achieving, which by the way, is off of a pretty strong year we had last year.
That's clearly a factor. We have also signed, agreements relative to securing capacity, and you can see some of those that have shown up in our 10-Q or 8-Ks we've issued, strategic partnerships we've entered into. Some of that has been, down payments and supply agreements to ensure that we have the right capacity. I think that's some part of it. Then it's hard to say, but on the last part of it, the sort of fungibility that may be coming into play a little bit, but it's too early to say, and quite frankly, the supply we're seeing today we can really mostly attribute that to sort of our own effort.
I will say, just to be very clear for everybody on the call, we're not out of the woods here. Our unfulfilled backlog has continued to grow each quarter. We're definitely working that down. We're focused very much on the end demand and shipping to the right people, and we feel like we've done a good job of that in terms of keeping our customers' lines moving and not causing impacts and being very agile. I think to the extent that on that third point, the fungibility side, if there were to be, a true, reduction or capitulation, let's call it, of the PC and consumer end market and not needing all the supply they've consumed, that certainly could be an incremental thing we could look at in the second half.
That's not quite there yet, but maybe some of it is. We do see pockets of things opening up, but it's really hard when you sort of drill into it to figure out where it exactly came from. Hope that's helpful.
Thanks, Matt.
Kind of a combination, but you sort of nailed it in terms of the different pieces.
Thanks, Matt.
The next question is from Timothy Arcuri with UBS. Please go ahead.
Hi. Thanks. Matt, I have just a broad question, kind of on-demand signals. To some degree because supply has been so tight, it does create the potential that some of the orders, and I'm not saying for you, but I'm just saying broadly, are being placed just to secure their spot in line and not really backed by underlying sell-through at the customer or even the customer's customer level. It sort of creates the risk that the chip companies are gonna see more of a delay in sort of the demand signals this time. I think it's less of a risk for you, but I guess my question is, how do you assess that as a risk?
Really the question, I guess, is on the tenor of your discussions with your customers. Do you get the sense that there are some examples of customers placing bookings just to get in the queue because you have such long lead times? I ask because some of the large cloud customers are beginning to slow some of their hiring, things like that. I'm just wondering what the tenor is of your talks with them. Thanks.
Sure, Tim. No, it's not unlike other cycles where certainly you're the phenomenon you point out, and this is the investor concern, right, more broadly in semis is have customers over-indexed on their order placements and what happens when demand stops or slows down. We've taken this very seriously. I mean, I just did pretty much half-day review in the last few weeks with my entire team, going through each of our product lines, going through the customers, and really kind of working through at an individual level where we still see the demand, do we see any inventory building up? If so, what does it look like? That type of exercise.
What I can tell you is there are no real examples I could point to Tim or anything thematic. I guess in some ways we may have the benefit that we've been somewhat behind, although I think we've done a good job of keeping everybody's lines moving. Certainly we haven't been the company that's been swimming in capacity. I think as we look out now, especially where the kind of global economy is beginning to moderate we're able to kind of take a look now and do a refresh, especially with our improved supply situation, and really make sure we're allocating it to the right people.
It's a very active process, Tim, but I can't give you any nuggets of specific examples where I think thematically inventory may be building, but I guess I would conclude by saying we were pretty transparent in the prepared remarks about what we saw in the consumer segment, which albeit i t's 12% of the business, and of that a smaller portion of it is really exposed to what I would call kind of true consumer like PCs and things like that. But that did very much slow down. There definitely was a change in demand, and as a result, some supply that had or some inventory that had built up. But we've got other moving pieces, even with consumer that's sort of offsetting it.
A long, long way of answering it. I don't have any great examples, but we're on it. We're watching it. But I'd say the bulk of the meeting, to be honest, when I had it, was the demand signals being very strong, right? Across the cloud companies, across our 5G business, enterprise, and wired. There are a lot of great opportunities for us, and a lot of this, Tim, is driven on Marvell's product cycles. It's not really necessarily an end market phenomena we're dealing with here. It's really the new product ramps that you're very aware of across these different segments, and those seem very much intact in aggregate.
Thanks, Matt. Yeah, I mean, I think in the continuum of companies that have risk, you're at the kind of bottom of the list. Thanks for the questions or the answers. Thanks.
The next question is from Vivek Arya with Bank of America Securities. Please go ahead.
Thank you for taking my question. Just wanted to clarify, Matt, if you have seen already or are baking in any supply or demand effects of the China lockdowns or Europe turmoil. My bigger question is, how do we conceptually align this notion of rising delinquency or unmet demand with rising inventory at, both on semiconductor company balance sheets and even at some of the large networking OEMs? What point does this rising inventory across the supply chain start to become a concern?
Great. Thanks, Vivek. Yeah. On the first question again, based on the timing of our reporting, we're like everybody else who's going right now. We had the month of April to comprehend in there relative to the China lockdown. We navigated it really well. I think one of the reasons for us is that when COVID first happened when the pandemic first began and there was major lockdowns everywhere, I think China including and also Southeast Asia, our operations team did a really good job of building up supply chain redundancy and resiliency over the last 2 years.
We've effectively dual-sourced almost everything we can, coming in and out of China and even some of the other regions over there. There was some impact. There was a little bit of product revenue where we could have gotten supply out of China, we could have fulfilled it. But it was very small, Vivek, to be honest with you. We're able to be fairly agile in terms of moving things around. That was a positive. On the war, in Ukraine and the impact on, kind of in the EU, again, I think you're hearing the data points.
We don't see it directly, but certainly on the slowdown in the consumer segment, people's budgets being pinched, things like that. I think that's part of the issue with some of the consumer end market commentary. We're not really in there. We don't have operations in Ukraine or Russia, so we didn't really have any impact on that. I think we're okay there. On the commentary on rising inventory, look, I'm very aware of this. You can pull everybody's 10-Q's out and look at one simple thing we'll do is look at OEM companies' revenue growth and OEM companies' inventory growth. You can see across a wide range of companies and industries,: growth in inventory far exceeding revenue growth.
That is something we're paying attention to. We're using that to dive in deep with our customers. I would say that our view tends to be like a lot of our customers, which is the end demand, especially in enterprise, in cloud, and in our product cycle in 5G, the demand looks pretty strong. There are missing pieces. There is the golden screw issue for sure that some of our customers have. We're just trying not to be the golden screw and also be very kind of clear-eyed about where the inventory is and what does that mean for Marvell.
All I can tell you is we're digging into it at a part level detail and making sure that we're being prudent about where we allocate the supply. It is something we gotta keep watching for sure.
Thank you.
To add to what Matt said is we actually chose to increase our inventory to support the revenue growth in second half. We have a very broad set of product lines, and we see strong demand. The flexibility we need to have raw materials and the wafers to support the revenue growth is really critical. We did choose that to make sure we can support the customers and the future revenue growth.
Do you have a target, Jean, on days of inventory?
The current level is what we're comfortable. We think we can help and manage the supplies for the second half. I think if the environment is normalized, we definitely, will have a lower inventory, like 120 days. That's what we think we should keep. Right now, we definitely chose to increase it to provide more flexibility.
Thanks very much.
Vivek, I just want to chime in on Jean's comment. I fully agree. If you look at just the steepness of the revenue ramp we're dealing with, clearly the whip has to go up, right? Just to sort of support the growth that part we're comfortable with. We'll watch it, and we'll certainly work it down, but I don't think you're gonna see many companies go back to the good old days of 80 days or something like that unless they truly get into trouble. I just think for a while, we're all gonna have to be mindful that there's a lot of volatility in the world.
We're sort of in the range now, and we're very conscious in terms of what we're doing here.
Thank you.
The next question is from Matt Ramsay with Cowen. Please go ahead.
Yes. Thank you very much. Good afternoon. Matt, in part of your script, you had talked about the upcoming server launches that include a couple of new technologies, right? The PCIe Gen 5 and then some CXL support for shared memory. I think you guys have done a great job of articulating what that might mean for your accelerator business and given us some outlook there. I was curious if you might expand a little bit about what PCIe 5 might drive for your support of that technology into the storage and network peripherals. Is that material to the growth in the back half, and how much visibility do you have there? Then just a quick follow-up. A question I get all the time is visibility on cloud CapEx.
If you guys look out in your cloud business, is it quarters, years? Like, how long are we talking that you feel like you have visibility 'cause there's definitely some concerns there. Appreciate it. Thank you.
Hey, Matt. Good to hear from you and good questions. I think you did nail it on one of the more important aspects of the upcoming CPU cycles is the I/O, right? The release of PCIe Gen 5. We've certainly been preparing for that. That is a key part of the equation for us, particularly as you can imagine in our storage business, in our data center storage business, which one of the key things we have to align to on our product roadmap is obviously nailing the PCIe I/O roadmap, and it's not a trivial thing, not only the controller, but obviously the PHY and getting everything qualified and working. Our team's done a great job there.
We're very well positioned in storage on PCIe Gen 5. I actually said in my remarks, we've already got wins lined up on Gen 6. That's still a ways out, but that's gonna be another leg of growth. You're right. The PCIe standard is really gonna be the fabric and the interconnect that CXL runs on top of. A lot of the effort we've put in on both Gen 5 and in the future, Gen 6, is gonna give us tremendous leverage as we think about disaggregation, particularly of memory and these different solutions that are gonna emerge around CXL. It's very strategic. We do.
You know, we always make or buy decisions on things in terms of make versus buy, sorry, on technologies, but this is one particularly on the PHY layer for PCIe that we do ourselves, and we're best in class in what we do here, and then we leverage that into our larger SoCs. But I think the biggest impact for Marvell will be just driving, broader adoption of our data center flash controllers, which again are well positioned across a number of NAND OEMs and products defined directly with the cloud OEMs, going back several years, right, to get all this lined up. It's gonna be a good cycle for us. Oh, and then, sorry, on cloud CapEx, I don't have any insight.
You know, we certainly look at all the publicly available data. You know, the most recent updates certainly look promising. I would say for Marvell, Matt, and for the rest of the investors on the call, we're aware, of course, of the overall CapEx environment and cloud can certainly be a tailwind, and it can be a headwind for anybody that supplies into there. You got to remember, most of our revenue coming from the cloud is really in new product cycles that are very Marvell specific in some cases. I mean, let's take PAM, for example. You know, if you went back a year ago, you really had a couple of companies deploying it versus older NRZ-based solutions.
You get into this year, there are now a dditional companies as well that are gonna be driving that cycle on 200 and 400 gig. We have new applications like even 800 gig as an example, for AI clusters and things like that, where we started ramping our 800 gig product. Whether the CapEx goes up or down, there's new applications and new replacement cycles that are going on. Then there's just other things like, for example, in our cloud Ethernet switch area, right? There was new wins we got through the Innovium team that joined us last year. We expect continued strong growth from that product area. We have new cloud-optimized silicon in ASICs and DPUs that were designed in over the last couple of years.
Those are gonna be ramping up. We are paying attention to it, Matt, but I would say that it's a little bit like our 5G business. You can see things go up and down and lumpiness in the market, but when you have a strong product cycle with either content or share gain and new customers coming on, you can power through. That's our full expectation in the cloud. We have very strong conviction on our business there today, particularly in the second half and next year, relative to the cloud hyperscalers. We'll watch the CapEx trend, but most of our products are independent of that trend.
Thanks, Matt. Appreciate all the color.
The next question is from Ross Seymore with Deutsche Bank. Please go ahead.
Hey, guys. Thanks for letting me ask a question. Jean, I wanted to ask you one on the margin side of the equation. You guys always run a very tight ship in this, but there are lots of moving parts. Matt's talked a lot about more supply coming on in the back half of this year. Sometimes that can come at a higher price these days. And then on even the OpEx side of things, wage inflation, et cetera, and some acquisition integration that you talked about. Generally speaking, can you walk us through the puts and takes on the gross margin and the OpEx side of things for the rest of the year?
Ross, thanks for the question. I think our team has been very disciplined to manage OpEx to make sure we follow our overall business model, right? Grow revenue significantly faster than OpEx growth and keep the gross margin right in the middle or even above the middle for our long-term target model, so we can drive the earnings expansion much faster than top line revenue growth. When you look at the OpEx Q1, as I said earlier, we increased headcount and project expense to really execute on all the design wins we have won. Q2, it's we keep it flat. Second half, it's going to go a little bit higher, slightly higher, but it's largely because we are going to grow revenue in second half. To support the revenue growth, you have certain variable expense that will increase slightly in second half.
In the longer term, you should continue to expect us to drive top line revenue growth and OpEx around 50% slower than top line revenue growth, so we can continue get earnings expansion. In Q2, if you look at the middle point of our guide, our operating margin is already about 36%. We are very driven to get our operating model to the 38%-40% range.
Thank you.
The next question is from Tore Svanberg with Stifel. Please go ahead.
Yes, thank you, and congratulations on the strong results. Just had a follow-up on the storage question, but this time on Gen 6, not on Gen 5. Just, Matt, can you maybe frame a little bit of Marvell's positioning here? I mean, I think about the PAM4 IP you have. Obviously, you talked about CXL. I know it's still early days, but if you could just frame a little bit layman's terms, your position for PCI Gen 6, that'd be great. Thank you.
Sure, Tore. Gen 6 as it relates to storage, you're absolutely right. Given the increase in I/O throughput and data rate, you had to move to a PAM-based architecture. That's played right into our strengths. We've coupled that with our 5-nanometer technology platform. We were early to do this, by the way. There's just a whole lot of goodness and reuse there, as you can imagine. It really put us in a strong position to get there early. I give our storage business unit a lot of credit for really having thought leadership here. I mean, they probably jumped, like, 2 nodes on everybody.
Not that it's a whole node game, but in terms of just the power and the performance that we're getting out of these solutions. Then, of course, the benefit we get to take the IO down to 5 nanometers and leveraging PAM is actually a huge help, right? Just given the complexity. You saw that in our optical business, right? Kind of how Inphi drove performance over time as they've driven PAM on new process nodes. Same thing over here. I think that's gonna be very strategic. I think that platform now. Again, we're early, right?
We're just talking about a Gen 5 launch today, but we do like to give color on our forward-looking roadmap. It definitely plays to our strengths. I think as you think about also what's gonna be required in CXL, I think even things like closing the link and the need for retimers and redrivers, accelerators pooling applications, they're all gonna require robust Gen 6 IP and then be able to integrate those into various different SoCs. I would just say even today, the CXL activity we're seeing in terms of our design opportunities is pretty broad and it's almost all, it's really, it's all in our 5-nanometer platform.
It all kind of hangs together, Tore, and maybe I think you sort of see this, but I think others are as well. The combination of all the IPs that we brought together through the different acquisitions, plus the Marvell technology platform, has really created an opportunity for us to go to market with these semi-custom solutions and be very successful in this next wave of cloud-optimized silicon.
Excellent. Thank you.
The next question is from Harlan Sur with J.P. Morgan. Please go ahead.
Good afternoon. Great job on the quarterly execution. I saw the announcement on, and you mentioned this, Matt, on shipping 100,000 400 gig coherent DSP chipsets. You're pretty early days in this transition, but you guys have already shipped, like, according to my calculations, almost $200 million plus worth of products. The great thing about this family of products is that they've got the same underlying architecture, and you've taken that, and you're going after different markets, right? Data center interconnect with 400 ZR, long-haul metro with your 400 gig Canopus solutions. The adoption seems to be actually on a faster curve. Is this a driver of the upside in your data center and carrier franchises? Then I believe that the ZR deployment for DCI is being driven by just one cloud titan.
Are you guys anticipating other cloud titans starting to hire as the year unfolds?
Thanks, Harlan, for the question. Maybe just to start at the top on coherent DSP. I'm not sure everybody quite understood, even the Inphi people that followed Inphi closely, how strategic this technology is. A little bit of background they had their own effort in coherent Inphi, and they ended up acquiring a company called ClariPhy a few years before we acquired Inphi. Ultimately, they chose the ClariPhy coherent DSP roadmap. This is very, very difficult stuff to do, Harlan. Once you get it right and get it designed in, these are very, very long product life cycles, particularly in the metro long-haul type of markets, the telco type markets.
That's really where the revenue is showing up today for this. The added benefit is obviously the same technology in chips are used in the ZR 400Z, the ZR market as well for DCI.
Yes.
There's a double benefit here. On the merchant side, we're a strong leader. We have a roadmap obviously beyond this to 800 gig. That's gonna play in well to our next generation DCI products as well. Even longer term there's discussions longer term about the modulation technology eventually inside data center moving to coherent.
It's a very strategic technology for Marvell. To answer your question on the ZR, DCI side, I said it in my prepared remarks, but it might have gotten lost. We do have multiple customers for this, not just one. Those will be ramping up over the next year or so, the additional customers. We're already shipping and putting out record results on the DCI products as it is. We're really driven by one customer, but there's more coming. So we're pretty excited about this. We have a whole ecosystem lined up, right? We have a number of module partners for different applications.
Anyway it's something we'll be talking about more. Maybe we'll do a tech talk on that one too, Harlan. It's actually really important technology that we got from Inphi, and it's obviously an outstanding team that really feeds a lot of the other things that we do inside the company.
Great insight. Thank you, Matt.
Just one final thing I would say, remember too, this is a chipset as well.
It's not just the DSP, it's the driver, it's the TIA, and there's even other components.
We're looking at to really optimize the solution. When you hang all that together and you can deliver it as a platform, it's very, very powerful.
Perfect. Thank you.
Welcome.
The final question today will be from C.J. Muse with Evercore. Please go ahead.
Good afternoon. Thank you for taking the question and squeezing me in. I guess a question on gross margins. You guided down 25 basis points sequentially, obviously a very small amount. Just curious if there's anything mix wise and/or higher input costs that we should be thinking about there. Just as a quick follow-up to Jean, are we at the net leverage level now where you plan to return free cash flow via buybacks, or is that still in more of the opportunistic phase? Thank you.
C.J., thanks for the question. First, on the net leverage, if you look at our Q2 guidance, at the middle point of our guidance, our net leverage ratio will be below 2. Gross debt to EBITDA ratio will be around the 2, but the net leverage will be below 2. As I said earlier, we have restarted share repurchase program, even after end of Q1. Subsequently, we did purchase additional $50 million through our 10b5-1 program. Our capital allocation principle is still the same, right? We want to invest, but we are also committed to return more than 50% free cash flow to shareholders, primarily through share repurchase. On the gross margin. Our gross margin, we're very pleased with the Q1 gross margin, which is record in the Q2 guide.
Both are above midpoint of our long-term target. The Q2 one is really just mix. We have so many different product lines quarter-over-quarter, you will see the mix change up and down around our midpoint. I would not predict that closely, but in general, it should be around the 65%, sometimes higher.
[Crosstalk]
I'll just add C.J., totally agree with Jean, obviously. Just to remind everybody, we set a range of on gross margins as our target model of 64%-66%. We had a record 65.3% in Q4. We had another record of 65.5% in Q1, and we're guiding 65.25% if you look at the midpoint in Q2. It is kinda in the range, if what I mean. I wouldn't read too much into it. I think running the company around that 65-ish is a really good place to be if we can continue to grow the company the way we're growing it and drive leverage and drive design wins and stay competitive.
I think we're in pretty good shape. I think obviously reasonable to ask the question, but I think as Jean said, we're gonna have little movements here and there on mix and we're ultimately just focused on keeping it in the right range and the right strike zone and then just driving really strong top-line growth and driving operating leverage through the cycle we're in here.
Thank you.
Yep.
This concludes our question and answer session, and the conference has also now concluded. Thank you for attending today's presentation. You may now disconnect.