Marvell Technology expects its data center business to account for more than 80% of total revenue in the near term, a sharp rise from less than 10% a decade ago, as artificial intelligence infrastructure demand accelerates. The company’s AI-related data center sales have already reached $2 billion per quarter, according to CEO Matt Murphy, speaking at the Six Five Summit in Santa Clara on Wednesday.
The chipmaker’s data center revenue expansion reflects broader industry trends, with cumulative data center capital spending projected to exceed $12 trillion through 2030, up from a prior estimate of $10.7 trillion. Murphy highlighted the growing total addressable market for AI infrastructure, noting that the cumulative XPU market opportunity could approach $700 billion between now and 2030. Marvell’s stock has surged 224% over the past year and 203% in the last six months, trading at $243.74 as of late Wednesday, up 1.40%.
Murphy emphasized that AI deployments remain in early stages, despite rapid adoption. He pointed to connectivity advancements, including CXL (Compute Express Link) and silicon photonics, as critical enablers for scaling AI workloads. The company’s silicon photonics portfolio has shipped millions of units and amassed 15 billion hours of reliability data over the past decade, Murphy noted. Marvell’s growth has been fueled by strategic acquisitions, including Inphi in 2021, Celestial AI in late 2025, and Avera Semiconductor in 2019, which bolstered its PAM-based DSP, chiplet photonics, and custom silicon design capabilities.
The company holds 15 to 18-plus design wins across the four major hyperscalers, positioning it to benefit from rising AI infrastructure investments. Murphy also referenced NVIDIA’s $2 billion investment in Marvell earlier this year, which included IP licensing provisions for custom product development. Gross profit margins stand at 51.5%, with a P/E ratio of 83 and a PEG ratio of 0.14. Murphy underscored the long-term potential of AI infrastructure, stating that the current technology ecosystem is still in its infancy and that connectivity innovations will unlock new use cases and cost efficiencies.
