Marvell Forecasts FY2031 Revenue to Reach Up to $90 Billion Driven by AI
At its investor day, Marvell Technology projected that its revenue could reach $70 billion to $90 billion in fiscal year 2031. This forecast represents nearly a tenfold increase over its expected FY2026 revenue of $8.195 billion, driven by massive demand for AI infrastructure. The aggressive forecast highlights the rapidly growing importance of high-speed interconnects and custom silicon (ASICs) in scaling AI data centers. Marvell estimates its total addressable market will expand to $400 billion by 2030 as networking interconnects become a primary performance bottleneck. Marvell raised its FY2028 revenue expectation by 20% to $20 billion, representing a 67% year-over-year growth rate, and expects its custom chip business to exceed $12 billion in revenue by FY2029. Overall, the company expects a compound annual growth rate (CAGR) of 55% to 60% between FY2026 and FY2031.
## BACKGROUND
Modern AI data centers rely heavily on high-bandwidth, low-latency interconnects and custom Application-Specific Integrated Circuits (ASICs) to transfer data across thousands of compute nodes. Unlike traditional computing, AI workloads generate massive continuous traffic across nodes, making optical and electrical networking hardware just as critical as raw GPU processing power.