Nvidia has invested US$2 billion in Marvell Technology as part of a strategic partnership designed to integrate Marvell’s custom chip designs with Nvidia’s proprietary networking and processing ecosystem. The deal, announced March 31, centers on the use of Nvidia NVLink Fusion, a rack-scale platform that allows customers to develop semi-custom AI infrastructure that remains fully compatible with Nvidia’s existing hardware and software stack.
The collaboration aims to address the growing demand for specialized AI hardware as major technology firms, including Alphabet and Meta, are projected to spend at least US$630 billion on AI infrastructure this year. Following the announcement, shares of Marvell rose approximately 7%, while Nvidia shares increased by 2.7%.
Integration and Technical Specifications
Through the NVLink Fusion platform, Marvell will provide custom “XPUs”—a term for various types of data processing units—and scale-up networking solutions. Nvidia will supply the foundational supporting technologies, including Vera central processing units (CPUs), ConnectX network interface cards (NICs), BlueField data processing units (DPUs), and Spectrum-X switches.
The partnership focuses on creating a heterogeneous AI infrastructure. This allows customers who design their own specialized processors to maintain seamless integration with Nvidia’s graphics processing units (GPUs), language processing units (LPUs), and storage platforms.
“The inference inflection has arrived. Token generation demand is surging, and the world is racing to build AI factories. Together with Marvell, we are enabling customers to leverage Nvidia’s AI infrastructure ecosystem and scale to build specialized AI compute,” says Jensen Huang, founder and CEO, Nvidia.
High-Speed Connectivity and Telecommunications
Beyond data center compute, the two companies will collaborate on silicon photonics and advanced optical interconnect solutions. These technologies are critical for high-speed, energy-efficient data transmission, which currently serves as a primary bottleneck in scaling large-scale AI systems.
The partnership also extends into the telecommunications sector. The companies intend to utilize the Nvidia Aerial AI-RAN (Radio Access Network) to transition 5G and 6G networks into AI-capable infrastructure.
“Our expanded partnership with Nvidia reflects the growing importance of high-speed connectivity, optical interconnect and accelerated infrastructure in scaling AI. By connecting Marvell’s leadership in high-performance analog, optical DSP, silicon photonics and custom silicon to Nvidia’s expanding AI ecosystem through NVLink Fusion, we are enabling customers to build scalable, efficient AI infrastructure,” says Matt Murphy, Chairman and CEO, Marvell.
Market Positioning and Financial Outlook
Industry analysts suggest the US$2 billion investment serves a dual purpose: securing Nvidia’s central role in data centers while accommodating the industry shift toward custom silicon.
“Nvidia gains access to Marvell’s semi-custom silicon and advanced optical interconnect capabilities to help scale data center-level AI systems where bandwidth and power efficiency are key bottlenecks,” said Jacob Bourne, Analyst, EMarketer. Bourne noted that the deal allows chips from other suppliers to operate within Nvidia-dominated environments
Marvell has previously indicated strong growth projections tied to the AI sector, stating it expects revenue to approach US$15 billion in fiscal 2028, representing a growth rate of nearly 40%.