Marvell Technology stock rallies on strong forward guidance for AI-focused custom silicon and accelerator products.
Marvell Technology (MRVL) shares surged 16% to $87.71 in early Friday trading, driven by intensifying demand for specialized AI hardware and the company's custom chips and interconnect solutions. This momentum reflects a broader industry shift toward expanded AI-chip expenditures, with Broadcom (AVGO) forecasting over $100 billion in AI chip revenues for next year, marking a significant expansion of hyperscaler investments beyond Nvidia (NVDA) dominance.
Marvell's strategic positioning in the AI ecosystem is strengthened by its collaborations with major cloud providers, including Amazon (AMZN) through AWS and Microsoft (MSFT), where it expects deeper integration of custom XPUs and enhanced optical components in data centers over the next five years. The company's president and COO, Chris Koopmans, emphasized that "the need for its custom ASICs and high-speed interconnects, which efficiently manage data flow among AI processors, memory, and servers, remains in massive expansion mode." These application-specific integrated circuits, designed for specific workloads, deliver superior efficiency compared to general-purpose GPUs, aligning with the escalating requirements of AI infrastructure.
Analysts at Citigroup (C) note that while Broadcom emphasizes scale-out co-packaged optics, Marvell is advancing scale-up designs with hyperscalers, potentially elevating its influence in future AI cluster architectures. Morningstar analyst William Kerwin anticipates this will drive further market share gains for Marvell in custom solutions. Supporting this optimism, Marvell's fiscal 2028 revenue is projected to increase nearly 40% to about $15 billion, surpassing the $12.92 billion consensus, while fiscal 2027 growth is revised to over 30%, approaching $11 billion.
The company's expansion is fueled by substantial capital investments in AI by leading technology firms, with Alphabet (GOOGL), Meta (META), Microsoft, and Amazon collectively projected to spend more than $630 billion this year on infrastructure. These outlays are accelerating adoption of Marvell's optical DSP chips, essential for high-speed connections within AI servers, alongside a faster-than-expected ramp in its custom AI processors. In the data-center segment, these developments position Marvell for sustained strength, building on its expertise in semiconductor design that optimizes power consumption and performance in hyperscale environments.
From a valuation perspective, Marvell currently trades at a 12-month forward price-to-earnings ratio of 20, which is lower than Broadcom's 25.30, suggesting potential upside as its AI-driven revenues materialize. This comparative metric reflects investor confidence in Marvell's ability to capitalize on the AI boom, where interconnect technologies play a critical role in enabling scalable, efficient computing clusters that support advancements in machine learning and data processing.