Marvell Technology, Inc. (MRVL.US) Fiscal Year 2026 Q4 Conference Call: Revenue Outlook for Fiscal Years 2027 and 2028 Raised Again.

Marvell Technology, Inc. (MRVL.US) held a conference call for the fourth quarter of the 2026 fiscal year. The company's total revenue for FY2027 is expected to approach $11 billion, with a year-on-year growth of over 30%, higher than the previous targets of $9.5 billion in September 2025 and $10 billion in December, which have been continuously raised. Data center revenue is expected to grow by 40% year-on-year, while communication and other businesses are expected to grow by 10%. Quarterly revenue is increasing each quarter, with FY4 2027 revenue expected to exceed $3 billion, with an increasing year-on-year growth rate. FY2028 total revenue is expected to reach around $15 billion, with a year-on-year growth of nearly 40%, an increase of approximately $2 billion from the December 2025 targets. Data center revenue is expected to increase by nearly 50%, while communication and other businesses are expected to have single-digit growth. Non-GAAP earnings per share will exceed $5. Celestial AI and XConn are expected to contribute approximately $2.5 billion in revenue in the 2028 fiscal year, as the acquisitions begin to deliver results. The company's executives stated that the growth in FY28 will be supported by three core DRIVE factors. Firstly, existing custom projects will continue to contribute to growth. Secondly, multiple XPU projects will achieve large-scale production, with particular success in customized NIC and CXL applications. Finally, the company's new flagship XPU project will enter large-scale production. The main XPU project is transitioning to the next generation and has secured orders for the entire year; the XPU project of the second top cloud provider is progressing well and is expected to be mass-produced in FY28. Demand for CXL extenders and custom NICs is accelerating, with these two applications alone expected to contribute over $2 billion in revenue by FY29. Q&A: Q: In addition to the growth rate of revenue, can you discuss the growth overview from the perspective of customer demand? The market has been concerned about the concentration risk of custom business customers, has the customer base expanded? A: Currently, we are deeply involved in the entire ecosystem, and we have a strong position among the four major US hyperscale cloud computing manufacturers and sub-tier manufacturers. Although the revenue share and product mix vary among different customers, it is important to note that, despite pushing the company's annual revenue target to $11 billion, the contribution of custom business to total revenue is not high, and it is not the main cause of customer concentration. In fact, as these four hyperscale players dominate a significant portion of the market's capital expenditure, it is natural for funds to flow towards them, which is an inevitable result of industry design. The diversity within these core customers is very high. As I mentioned before with a long list of product offerings, we provide different and rich product mixes for these four customers. Although the custom business has attracted a lot of attention, it is just one part of the overall equation. Looking ahead to the 2028 and 2029 fiscal years, as more than 20 mature or soon-to-be in-production design projects we hold gradually start releasing, the customer diversity within the custom business will further increase. Our unique advantage lies in the depth and breadth of our product line, which is sufficient to meet the needs of core supercomputing customers end-to-end. Our recent two acquisitions perfectly complement our capabilities in PCIe, UAL, and key silicon photonics technologies, further strengthening our service boundaries with these top customers. Q: OpenAI recently formed a partnership with your leading XPU customer, expecting to consume about 2GW of next-generation XPU capacity. This shows that AI computing demand is still accelerating. Considering that your company will have 15-20 XPU accommodation custom projects coming to light in the next two years, and the next-generation XPU project of the leading customer has started to climb, would you say that this project will continue to have stronger stepwise growth in the second half of the year, or will it trend linearly? Has there been any change in the previously expected annualized revenue of $2 billion? A: The strong validation of AI computing expenditure in the market, especially the trend of large companies continuously investing in developing their own XPUs, is a very positive signal for us. Even in some customers where we are not directly involved in XPU design, our XPU auxiliary products penetrate comprehensively. This high level of participation by across large internet companies supports our judgement that the custom business will grow beyond expectations this year, and creates a foundation for explosive growth in the years 2028 to 2029. In terms of growth pace, as a result of project conversions, the expectation of stronger growth in the custom business in the second half of the year remains valid. Regarding the annualized revenue you mentioned, not only will the previous target remain unchanged, there is even room for an upward adjustment. The current overall end-of-year revenue target of the company is already above $3 billion (Q4 single quarter), with a significant potential incremental contribution from custom business. As the projects launched in the second half of the year enter into a full year's contribution period next year, FY28 revenue will experience a substantial leap forward. The growth momentum mainly comes from three factors: the increase in product value and volume, the tiered introduction of XPU auxiliary products, and the start of cooperative projects with new top-tier large-scale customers. The initial forecasts for this new project are very conservative, but in reality, our wafer production and production planning with the supply chain are far higher than the current reported forecast values. If the current trend continues, the performance next year has a great potential for upward revisions. Based on past experience, we are used to giving conservative forecasts first, gradually revising them as projects progress, so achieving double-digit growth in the custom business next year is not a radical assumption. Q: Can you clarify the growth base of the custom business in the 2026 calendar year (FY2027)? Will the previously expected 20% growth rate be raised to around 30%? Also, with regard to the XPU project of the second major customer, how confident is the company in the timing of mass production and the "doubling expansion" target? A: With regards to the guidance for the 2027 fiscal year, based on the doubling growth of the custom business, we estimated a growth rate of 20% this year, and the current judgement is that the actual growth rate will be higher than this number. Although I cannot provide an exact revised figure at the moment, the expectation is clearly tilting upwards, and it is suggested to adjust upwards based on a 20% foundation, which will also result in a larger base for next year than originally anticipated. We are very confident in achieving our growth targets, primarily driven by Marvell's successful history of climbing in large-scale custom project development and production. We have clear anticipation of each key milestone in the introduction of new products, and have conducted in-depth alignment with customers on production plans. In fact, the capacity we have reserved for FY28 is far higher than the expected values revealed to the market. The current financial budget has already taken into account potential project delays and uncertainties, and there are still some doubts in the market about the landing of this project, so the forecasts we provide are very practical and reasonable. If the current trend continues, there is great potential for performance to be significantly revised upwards. According to past experiences, we are used to giving cautious forecasts first, gradually revising them as projects progress, so the realization of double-digit growth in the custom business next year is not an aggressive assumption. Q: Since the company raised its expectations for the 2028 fiscal year by $20 billion since December last year, what are the specific drivers behind this significant upward revision? What positive changes have occurred? Also, can this growth momentum continue through the 2028 calendar year? A: This is actually a process of continuously clarifying our understanding over time. Looking back to September last year, we had a revenue expectation of around $9.4 billion, which was raised to $10 billion in December, and is now pointing to over $11 billion. This progressive upward revision comes from clearer visibility and more specific project implementations. The most significant change is in the interconnect business. Frankly, our previous forecasts were very conservative, anchoring the growth rate merely to the overall capital expenditure of cloud providers. However, as some analysts have pointed out, this business should be more linked to GPU and XPU demands. The current order situation and backlog orders confirm this, leading to a significant increase in revenue forecast for the interconnect business, which then extends to the prospects for next year. What supports this is extremely strong order flow and in-depth communications we have had with customers regarding supply planning. As early as April 2024, we set a goal for the 2028 calendar year, assuming a 20% market share in the data center market, and many thought that the target of $15 billion in data center annual revenue was a pipe dream. However, at the AI Investor Day in June last year, with the expansion of the TAM, this target value had implicitly risen to over $18 billion. Looking at the current implementation situation, we are steadily realizing these ambitious goals that seemed "unattainable" at the time. This is not just an increase in figures but also a powerful validation of the strategic planning we made four years ago. While we still need to continue working hard at the execution level, the current order forecasts, forecasts, and synergy with the supply chain give us confidence in this year, the next year, and even the 2028 calendar year growth. Q: The management mentioned in the opening remarks that the AEC and Retimer business will double in the current fiscal year. Can you clarify the magnitude of this business? Additionally, what role do these emerging products play in the company's end-to-end interconnect strategy, and how do they contribute to the next few years? A: This is still a relatively new area for us. Although we expect growth to double this year, the current revenue scale is around $200 million. Based on current observations, this number still has further upward potential. In the semiconductor field, once such emerging products start to double, they often generate sustained inertia. We have great confidence in the competitiveness of our products in this area, especially in terms of leading technologies. We have made full use of DSP and PAM4 technologies in this field. The company made a strategic decision earlier to enter the Retimer market at a key juncture in the transition from NRZ to PAM encoding. Although we entered certain slots later than some existing manufacturers, we are in the early stages of the cycle and plan to make large-scale investments in-depth. In the long term, electrical interconnect and optical interconnect are complementary. While we have focused heavily on pluggable optical modules since the acquisition of Inphi five years ago, and more recently with the acquisition of Celestial AI for CPO technology, electrical interconnect is indispensable in specific distance and application scenarios. Our goal is to be a one-stop end-to-end connectivity partner for customers - whether in electrical transmission, optical transmission, silicon photonics technology, or different transmission distances and forms. What customers need is a highly reusable and reliable partner in IP, firmware, software, and system implementation. Although AEC and Retimer currently have a relatively small size within the company's overall picture, they are in a phase of rapid growth and represent the final missing piece in complementing our full-stack interconnect capabilities. Q: Could you update the product mix situation in the photonics business? Although 1.6T has started shipping, the market generally believes that 800G is still mainstream this year. Could you provide an estimate of how the distribution of 1.6T, 800G, and 400G may look in the 2027 fiscal year? A: Your assessment is generally accurate. We have always emphasized that 800G products will have a long lifecycle, and their strong momentum will persist through this year and the next, dominating a major portion of shipments. However, as I mentioned in the opening remarks, the first shipments of 1.6T products have shown significant progress towards climbing this year and are about to steeply increase. At the moment, one of the driving forces behind the increased outlook on the interconnect business for the entire year comes from increasing demands from all customers, with 1.6T growth particularly noticeable. The production demand from the initial customers is very strong, with more customers expected to join in the next two years. Due to the highly dynamic and dynamic order and demand environment, we cannot currently provide an exact breakdown of proportions. As this year's business unfolds further, the product mix profile will become clearer. Q: With such strong growth expectations, are there potential challenges in supply chain management? How has the company progressed in supply chain response compared to a few years ago? A: Since the release of ChatGPT, any high-end wafer manufacturing, advanced packaging, and large substrate areas involving AI have consistently been in a tight-supply balance. However, against this backdrop, we still achieved over 40% total revenue growth last year, which is proof of our extremely stable cooperation with suppliers. I believe the core advantage lies in our early anticipation and communication of this growth. We have laid a solid groundwork in anticipation of this round of growth, by providing suppliers with future demand visibility over the next several years. Based on the current resource dispositions, I am fully confident in securing the necessary supply quantity to support the growth blueprint we have outlined for this year, the next, and beyond. Q: The company has raised its revenue guidance to over $15 billion for next year, with an earnings per share (EPS) guidance of around $5. This revenue scale is approximately 15% higher than the current market consensus, but the increase in profit margin seems to only be half of that of the revenue side. Could you break down the changes below the revenue line? Is the gross margin affected by the product mix, or has there been an increased investment in research and development to achieve growth? A: About the EPS guidance of "over $5," this is more of a baseline reference rather than an exact forecast. You can calculate it based on the company's current financial framework. From the revenue side, we have provided a macro framework from which you can model based on the company's year-end revenue exceeding $3 billion. On the profit side, if you consider the operating expenditure guidance we have given and the trend in the gross margin, you will find that by the year-end of this year, we are already nearing the target operating model. A cautious assumption is that this level of margin performance will continue into next year. Based on a revenue foundation of $15 billion, coupled with the current profit margin model, the resulting EPS is expected to be above $5 for next year. I am not implying any deterioration in profit margins, earnings dilution, or loss of operating leverage. In fact, we are currently in a period of releasing operating leverage - the operating profit margin is in the range of about 35% and will continue climbing by the end of this year. Next year, the profit margin performance will at least be consistent with the year-end level of this year. In simple terms, based on these assumptions, the final profit figure will definitely be higher than $5.