Nvidia and Marvell Forge $2B Partnership for AI Infrastructure

Nvidia and Marvell have entered into a strategic partnership valued at $2 billion, aimed at deepening the integration of Marvell's custom silicon and interconnect portfolio into Nvidia's NVLink Fusion ecosystem. This collaboration positions Marvell as a key enabler of semi-custom AI infrastructure built on Nvidia architectures, reinforcing Nvidia's strategy to extend its platform beyond GPUs into a broader, partner-driven "AI factory" model. The agreement will see Marvell contribute custom XPUs and NVLink-compatible scale-up networking, while Nvidia provides the underlying compute and connectivity stack, including its Vera CPU, ConnectX NICs, BlueField DPUs, NVLink interconnect, and Spectrum-X switching platform. This partnership enables customers to integrate custom accelerators alongside Nvidia GPUs and other infrastructure components without compromising compatibility with Nvidia’s broader ecosystem, addressing the growing demand for specialized AI deployments and heterogeneous AI architectures. The companies will also collaborate on silicon photonics and optical interconnect technologies, and advance AI-RAN deployments for 5G and emerging 6G networks.

Counterparts (2)

Partner

Marvell

company

Marvell Technology, Inc. is a global semiconductor company that designs, develops, and markets a range of data storage, processing, networking, security and connectivity solutions. Their products are used in data centers, enterprise, automotive, and consumer markets.

NVIDIA

company · 5 deals

NVIDIA is a technology company specializing in accelerated computing, particularly in AI and digital twins. Their technologies are used across various industries, including automotive, healthcare, and manufacturing, with a focus on enabling sustainable computing through energy-efficient solutions.