What's happening
Marvell Technology entered discussions with Google to co-develop two custom AI inference chips, according to reports from The Information on April 19, 2026. The partnership would focus on creating a memory processing unit designed to work with Google's existing Tensor Processing Units (TPUs) and developing an entirely new TPU architecture optimized for AI inference tasks. Marvell shares jumped 6% in premarket trading on April 20, 2026, following the news, potentially adding more than $7 billion to the company's $122.15 billion market capitalization if gains held through regular trading hours.
Why it matters for markets
The potential Google partnership represents a significant revenue opportunity for Marvell, which expects its revenue to approach $15 billion in fiscal 2028. Custom AI chip partnerships have proven lucrative in the sector, with Meta paying Broadcom $2.3 billion last year for AI chip design and related services. The deal would also strengthen Marvell's position in AI hardware following Nvidia's $2 billion investment in the company in March 2026 to integrate custom AI chips with Nvidia's networking gear and processors. Marvell currently trades at 33.35 times next 12-month earnings estimates, a premium to Broadcom's 27.84 times multiple, reflecting investor expectations for growth in AI-related revenue streams. The partnership underscores the intensifying competition for AI inference workloads, where specialized chips can offer performance and cost advantages over general-purpose processors.
Sectors and assets to watch
The semiconductor sector faces continued consolidation around AI chip capabilities, with Marvell Technology (MRVL) positioned to benefit from custom chip partnerships alongside its existing $8.19 billion revenue base. Nvidia Corporation (NVDA), with its $4.90 trillion market capitalization and dominant position in AI training chips, faces growing competition in the inference market where specialized designs can challenge its H100 and Blackwell GPU architectures. Alphabet Inc. (GOOGL), with its $4.13 trillion market cap, continues diversifying its AI hardware supply chain beyond internal development, following the strategy of other hyperscalers seeking custom silicon solutions.
What to watch next
Monitor whether Marvell and Google formalize their partnership discussions into a definitive agreement, which could provide concrete revenue projections and timeline details. Track competitive responses from other semiconductor companies pursuing custom AI chip partnerships, particularly as hyperscalers seek alternatives to Nvidia's dominant position in AI hardware. Watch for updates on Marvell's progress toward its fiscal 2028 revenue target of $15 billion and how AI-related contracts contribute to that growth trajectory.