Marvell Set To Double AI Chip Sales To $4 Billion Next Year, Analyst Projects

JPMorgan analyst Harlan Sur projects Marvell Technology's AI-related revenue to more than double to $4 billion in 2025, driven by key partnerships with Amazon (Trainium 3) and Microsoft (MAIA Gen 3), alongside strong demand for optical DSPs from Nvidia and Alphabet. The company anticipates Q2 revenue of approximately $2.05 billion and Q3 guidance of $2.1 billion, with a recent automotive Ethernet business sale expected to be accretive to earnings through potential share buybacks despite a near-term EPS hit. Sur maintains a positive long-term outlook for Marvell, citing its strong position in hyperscale capex and AI infrastructure investments.

Analysis

According to a JPMorgan analyst, Marvell Technology's AI-related revenue is projected to more than double to $4 billion in 2025 and continue accelerating into 2026, positioning the company as a key beneficiary of the AI infrastructure build-out. This growth is anchored by two primary drivers: a strong pipeline of custom AI chips for major hyperscalers, including the production ramp of Amazon's Trainium 2 and a clear roadmap for Trainium 3 (3nm) and Microsoft's MAIA Gen 3 (2nm/3nm); and robust demand for its 800G PAM4 optical digital signal processors (DSPs) which are critical for interconnecting Nvidia's H200/GB200 GPUs and Google's TPU infrastructure. While near-term revenue guidance for the October quarter of approximately $2.1 billion is slightly below consensus, this is attributed to the strategic divestiture of its automotive Ethernet business. This sale is expected to cause a minor near-term EPS reduction of $0.02–$0.03 per quarter, but is projected to be accretive by $0.05–$0.10 per share if the proceeds are used for share repurchases, signaling a strategic pivot to higher-growth segments.

Key Decisions for Investors

• Investors with a long-term horizon should consider the multi-year AI revenue growth trajectory, underpinned by secured design wins with major cloud providers like Amazon and Microsoft.
• Monitor the company's execution on key product milestones, particularly the high-volume ramps for Amazon's Trainium 3 in 2026 and Microsoft's MAIA Gen 3 in 2027-2028, as these are critical to achieving the forecasted revenue acceleration.
• Closely track management's announcements regarding capital allocation following the automotive business divestiture, as the execution of a share buyback program would be a key catalyst for the guided EPS accretion.