NVIDIA Invests $2Billion in Marvell to Expand AI Infrastructure

NVIDIA commits $2B to Marvell Technology in a landmark AI infrastructure deal powered by NVLink Fusion, custom XPUs, and silicon photonics.

NVIDIA has announced a $2 billion investment in Marvell Technology, paired with a sweeping strategic partnership and AI-RAN ecosystem through its NVLink Fusion interconnect platform.

The announcement, made on March 31, 2026, marks one of Nvidia’s largest single ecosystem investments and signals a significant shift toward modular, semi-custom AI infrastructure.

Under the terms of the agreement, Marvell will provide custom XPUs and NVLink Fusion-compatible scale-up networking that can be integrated with NVIDIA hardware, including Vera CPUs, ConnectX NICs, Bluefield DPUs, NVLink interconnect, Spectrum-X switches, and rack-scale AI compute platforms.

NVLink Fusion, first announced in May 2025, enables heterogeneous AI infrastructure where non-Nvidia accelerators can communicate with GPUs, CPUs, and networking hardware over NVLink’s high-bandwidth, low-latency fabric. Platforms built through the program must include at least one product, whether a CPU, GPU, or switch.

Also Read: L&T, NVIDIA Partner to Build Gigawatt-Scale Sovereign AI Factory

The two companies also agreed to collaborate on the development of silicon photonics, which uses light instead of traditional copper wiring to move data faster and more efficiently.

Marvell’s acquisition of Celestial AI late last year added photonic fabric technology to its portfolio, and this deal now places that capability squarely inside it’s ecosystem. The two companies will also collaborate on AI-RAN infrastructure for 5G and 6G networks.

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.
Jensen Huang, CEO, Nvidia

Marvell Chairman and CEO Matt Murphy Growing importance of high-speed connectivity, optical interconnect, and accelerated infrastructure in scaling AI.
Matt Murphy, CEO, Marvell Chairman

Marvell shares rocketed over 10% in premarket trading following the announcement, underscoring the market’s enthusiastic reception. Earlier buzz from Marvell’s Q4 earnings beat with both revenue and EPS topping estimates added further positive momentum.

Nvidia has been deploying $2 billion checks across the AI stack, having backed companies including Synopsys, CoreWeave, Coherent, Lumentum, and Nebius Group, each investment tying another piece of the ecosystem closer to Nvidia’s platform.

Analysts at Wedbush noted that while the partnership involves NVLink, AI-RAN, and core technologies such as optical networking and silicon photonics, it is not entirely clear which technology is the primary driver for Nvidia. Wedbush highlighted the financial commitment itself as a major factor, suggesting Marvell was strongly motivated to finalise the deal given the scale of Nvidia’s backing.

This deal is not about competition, it’s about consolidation. It is methodically locking key semiconductor and networking players into its platform, making NVLink Fusion the gravitational center of next-generation AI compute. For Marvell, joining it’s ecosystem with a $2 billion vote of confidence signals a clear path toward becoming a cornerstone of the AI infrastructure supply chain.

Disclaimer : This article is intended for informational purposes only, and is based on publicly available information. NervNow has not independently verified the claims.

MORE ON NVIDIA:
NVIDIA, Emerald AI Partner on Flexible AI Factories as Grid Assets
LTM Wins NVIDIA Partner Network Rising Star Award at GTC 2026
Nvidia’s New AI Supergroup Has 8 Players, India’s Sarvam Is One of Them