Marvell’s AI momentum reignites the question: could it become “the next Nvidia”?

Marvell Technology has become one of the most closely watched semiconductors tied to the buildout of AI data centers, after its shares gained about 270% so far in 2026. The comparison to Nvidia, the company that dominated investor sentiment around AI compute over the past few years, has returned as Wall Street looks beyond GPUs to the networking and “data movement” layer that connects thousands of accelerators in modern AI systems. (Jun. 8, 2026, The Motley Fool)

Marvell’s pitch, and the reason it is getting pulled into “next Nvidia” conversations, is its focus on custom silicon and interconnect. The company sells application-specific integrated circuits, or ASICs, commonly referred to in its market language as XPUs, designed for specific computing tasks. It also sells Ethernet switching and optical digital announcement processing, technologies intended to reduce bottlenecks in latency, bandwidth, and power efficiency when training large models requires large-scale communication across server racks. (Jun. 8, 2026, The Motley Fool)

A key catalyst for the current run-up is Nvidia’s explicit support. In late March, Nvidia and Marvell announced a strategic partnership centered on Nvidia NVLink Fusion, a rack-scale platform meant to let “semi-custom” AI infrastructure plug into Nvidia’s NVLink ecosystem. Under the deal, Nvidia invested $2 billion in Marvell, while Marvell is positioned to provide custom XPUs and NVLink Fusion-compatible scale-up networking. Nvidia, in turn, would provide supporting components including its Vera CPU, ConnectX NICs, Bluefield DPUs, NVLink interconnect, and Spectrum-X switches, plus the rack-scale AI compute. The companies also said they will collaborate on silicon photonics technology. (Mar. 31, 2026, Nvidia)

For investors, the partnership matters because it indicates that Marvell’s chips are being designed to work in an Nvidia-centered AI architecture, rather than as an “alternative ecosystem” operating entirely on its own. Reporting around the time of the Computex show in early June also highlighted Nvidia CEO Jensen Huang’s public endorsement of Marvell as a next-generation standout, adding to the market narrative that Marvell could scale into a major AI-infrastructure platform. (Jun. 2, 2026, Axios)

The financial backdrop that investors are using to underwrite the story is Marvell’s recent results. In its fiscal year 2026 earnings release, Marvell reported record net revenue of $8.195 billion, up 42% year over year, attributing the growth to robust AI demand. In the fourth quarter, Marvell said net revenue hit $2.219 billion, also a record, and reported GAAP net income of $396.1 million. Just as important for the AI narrative, Marvell disclosed that data center was its largest end market, representing 74% of total net revenue in the quarter, with data center revenue of $1.6513 billion. (Marvell investor relations)

Still, the “next Nvidia” label runs into scale and business-model differences. Nvidia built a dominant position by supplying the bulk of AI training and inference compute, while Marvell is largely an enabler for the layers around that compute. The Motley Fool’s analysis argued that even if Marvell continues to benefit from the AI infrastructure boom, replicating Nvidia’s dominance would require Marvell to capture broader parts of the compute stack, which the competitive chip landscape makes difficult. (Jun. 8, 2026, The Motley Fool)

What remains uncertain is how much of Marvell’s current enthusiasm is already reflected in valuation, and whether partnership-driven designs will translate into sustained margin and growth improvements beyond the quarters already supported by AI demand. Marvell’s earnings release included upbeat commentary, including expectations for year-over-year revenue growth to accelerate in each quarter of fiscal 2027 and a “record pace” in bookings, but it did not break out performance specifically tied to NVLink Fusion deployments or quantify outcomes from silicon photonics collaboration. Investors will likely focus on upcoming earnings and disclosure around design wins, customer ramps, and whether Marvell’s AI networking and custom-silicon revenue mix continues to widen. (Marvell investor relations)