CHONGQING, CHINA – AUGUST 26: In this photo illustration, a smartphone displays the logo of Marvell Technology, Inc. (NASDAQ: MRVL), the American semiconductor company, with the company’s latest stock market chart shown in the background on August 26, 2025 in Chongqing, China. (Photo by Cheng Xin/Getty Images)
Getty Images
Marvell Technology (NASDAQ:MRVL) recently reported solid second-quarter results; however, a cautious outlook regarding its data center division caused the stock to decrease nearly 12% in after-hours trading on Wednesday. This decline compounds a 30% fall earlier this year, as investors express doubts about the company’s position in the rapidly evolving semiconductor and AI sectors. Nevertheless, the recent downturn has positioned Marvell as a potentially attractive AI investment. The stock now trades at approximately 25 times estimated fiscal 2026 earnings and 20 times fiscal 2027 earnings. This valuation seems justifiable given the company’s robust growth, with consensus projections indicating a 43% revenue increase for FY’26 and a 19% rise for FY’27. However, it’s essential to note that appealing valuations do not guarantee stability. The main concern is whether these figures can be relied upon, given the fluctuation in customer orders and the competitive landscape in AI hardware. To determine if Marvell truly presents an enticing entry point for investors, let’s examine the risks associated with its data center business and the prospects offered by its growing AI portfolio.
Some Headwinds
Marvell’s second-quarter results indicated relatively impressive growth. Net revenue reached a record $2.0 billion, representing a 58% increase compared to the same time last year. Non-GAAP net income was $585.5 million, or $0.67 per diluted share. Data center revenue climbed 69% to $1.49 billion, although it fell slightly short of analyst expectations. In spite of this strength, management has guided for flat sequential revenue from the data center sector in the third quarter, anticipating a rebound in the fourth quarter. The weak outlook reflects some worries about Marvell’s custom ASIC operations. Management highlighted “lumpiness” in orders from hyperscaler customers, indicating the unpredictable nature of demand.
Trends that may be influencing this situation include delays in Microsoft (NASDAQ:MSFT) next-generation AI chips. The Maia 200 has been rescheduled to 2026, while Braga-R and Clea have been postponed to 2028 or beyond due to design modifications and other challenges. This has raised concerns for Marvell Technology, which provides essential components for these chips. Moreover, Amazon’s (NASDAQ:AMZN)Amazon Web Services (AWS) has been ceding market share to Microsoft Azure and Google Cloud. Although AWS remains the largest cloud infrastructure provider by market share, its growth rate has slowed in comparison to Microsoft Azure and Google Cloud. This has created further uncertainty regarding Marvell’s order pipeline, as AWS also depends on Marvell for its Trainium AI chips. Should You Buy MSFT at $500?
Marvell’s Opportunity
Marvell’s entry into the AI market began with its high-speed interconnect solutions for data centers. These optical and electrical interconnects are vital for handling the vast amounts of data produced by AI and machine learning tasks. Given that these tasks depend on extensive parallel processing and rapid data transfer, they impose significant demands on the current infrastructure, thereby creating a requirement for the advanced connectivity technologies that Marvell delivers. The larger growth engine has been application-specific integrated circuits (ASICs) designed for AI. Unlike general-purpose GPUs, ASICs are tailored to the specifications of individual clients, such as hyperscalers, offering improved cost efficiency, reduced power consumption, and enhanced performance. This could serve Marvell well in the long run.
Major technology firms are in the midst of an extraordinary AI investment surge. Amazon is anticipated to allocate up to $105 billion on capital expenditures in 2025, with Microsoft, Alphabet, and Meta projected to spend as much as $80 billion, $75 billion, and $72 billion, respectively, a substantial portion of which is designated for AI infrastructure, including GPUs from Nvidia (NASDAQ:NVDA). However, investors will eventually start to place a premium on returns from AI investments, and companies may become more discerning regarding spending.
Large enterprises may look for alternatives to Nvidia, and Marvell’s specialized models might become a leading option for hyperscalers. Furthermore, as AI models increase in size, incremental performance enhancements may decline, and the availability of high-quality training data could present a limitation. The market might potentially transition from large-scale general-purpose AI models to smaller, specialized ones, benefiting more niche companies like Marvell that provide customized products designed to optimize costs and performance for distinct applications.
Returning to the stock, Marvell’s relative valuation of around 25 times forward earnings seems reasonable in comparison to Nvidia, which trades at approximately 40 times fiscal 2026 earnings. Certainly, Nvidia holds a more entrenched position in the AI market, and its revenue is expected to increase by more than 50% this year. However, Marvell may still present investors with a more affordable way to invest in AI infrastructure, especially as the demand for tailored, power-efficient solutions continues to rise.