NVIDIA is set to invest $2 billion in Silicon Valley chipmaker Marvell Technology, aiming to enhance networking technology and integrate its AI chips into larger-scale data centers.
The two semiconductor companies announced on Tuesday that they will collaborate on developing silicon photonics technology, utilizing optical advancements to upgrade data center network systems and improve data transmission speeds.
The deal between NVIDIA and Marvell will also assist major technology firms in connecting their custom-designed AI chips to NVIDIA's data center systems.
Marvell designs specialized AI accelerator chips for U.S. tech giants such as Amazon.com, which can serve as alternatives to NVIDIA's general-purpose Graphics Processing Units (GPUs).
The companies stated that this partnership will enable seamless integration of such custom AI chips with NVIDIA's GPUs, networking, and storage systems.
NVIDIA is leveraging its leadership in AI processors to expand into a comprehensive platform for AI data centers, solidifying its position in the core infrastructure where tech giants are investing billions to compete for AI dominance.
Earlier this month, NVIDIA introduced its first dedicated chip designed specifically for AI inference—the process by which AI models respond to user queries and generate outputs like text or code—marking the initial expansion of its core AI chip products beyond GPUs.
On Tuesday, Marvell's New York-listed shares surged 12.8%; over the past 12 months, the stock has gained more than 50%, bringing its market capitalization to $86.6 billion.
Marvell CEO Matt Murphy commented that the collaboration will help customers improve scalability and efficiency: "Our deepened partnership with NVIDIA underscores the growing importance of high-speed connectivity, optical interconnects, and accelerated infrastructure in scaling AI."
Both NVIDIA and Marvell have recently pursued acquisitions to strengthen their positions in the AI data center market.
Last month, Marvell completed a $3.3 billion acquisition of Celestial AI, which develops photonic technology capable of connecting hundreds of thousands of AI accelerator chips.
Clustering AI chips on a massive scale has become crucial for building and deploying cutting-edge AI systems, such as Google's Gemini, Anthropic's Claude Code, and OpenAI's ChatGPT.
Since acquiring networking equipment provider Mellanox for $6.9 billion in 2019, NVIDIA has gradually expanded its business beyond GPUs.
In December of last year, NVIDIA entered into an approximately $20 billion deal with AI chip startup Groq, absorbing its key talent and licensing its technology, laying the groundwork for the launch of a new inference chip this past March.
Comments