Nvidia Invests $2 Billion in Marvell to Strengthen AI Ecosystem

Nvidia has invested $2 billion in Marvell Technology as part of a broader strategy to maintain its central role in the rapidly evolving artificial intelligence infrastructure market.

The investment is designed to improve compatibility between Marvell’s semi-custom AI chips and Nvidia’s ecosystem, including its networking technologies and processors. As more companies explore custom silicon to reduce reliance on Nvidia’s high-cost GPUs, the move helps ensure those alternatives still integrate within Nvidia-dominated data center environments.

Marvell specializes in custom chip design and advanced networking solutions, particularly in optical interconnects and silicon photonics — technologies critical for high-speed, energy-efficient data transfer in large-scale AI systems.

Through this partnership, Marvell will provide custom silicon and networking components compatible with Nvidia’s NVLink Fusion architecture, while Nvidia will supply CPUs, network interface cards and interconnect technologies.

The deal reflects a strategic shift: rather than competing directly with all custom chip providers, Nvidia is positioning itself as the foundational platform enabling diverse AI hardware ecosystems.

Major technology firms such as Alphabet and Meta Platforms are expected to collectively spend over $600 billion on AI infrastructure this year, significantly boosting demand for advanced semiconductors and networking hardware.

Following the announcement, Marvell shares rose around 7%, while Nvidia also recorded gains, signaling strong investor confidence in the partnership.