NVIDIA is quietly stepping up its push into AI infrastructure, with reports suggesting a $2 billion investment in Marvell Technology. The move signals a deeper collaboration between the two companies at a time when demand for AI systems is rising sharply.

At its core, the partnership is about making AI infrastructure more efficient and better connected. By bringing Marvell’s networking and chip capabilities into NVIDIA’s ecosystem, both sides are aiming to improve how large-scale AI systems are built and operated.

One of the key elements in this collaboration is NVIDIA’s NVLink Fusion platform. It allows companies to design semi-custom AI setups by combining GPUs, networking, and storage in a more flexible way. With Marvell now part of this setup, the integration between hardware components is expected to become smoother and more scalable.

The partnership also looks beyond current infrastructure needs. Both companies plan to work on silicon photonics and telecom technologies using NVIDIA’s AI-RAN platform, which supports AI-powered 5G and future 6G networks. The idea is to develop faster optical connections that can handle the growing pressure of distributed AI workloads.

Marvell CEO Matt Murphy noted that as AI systems expand, fast connectivity and advanced interconnects are becoming just as critical as compute power itself.

The timing is notable. Competition in the semiconductor space is heating up, with companies like Samsung Electronics reportedly preparing to mass-produce silicon photonics chips by 2028, in a bid to compete with TSMC and address data bottlenecks in AI.

Under this partnership, Marvell will contribute custom processing units and networking technologies, while NVIDIA will provide its expertise in GPUs, interconnects, and data centre systems. Together, they aim to offer businesses more flexibility when building next-generation AI infrastructure.

NVIDIA CEO Jensen Huang also pointed to the rapid rise in AI usage, particularly in inference, where demand is accelerating quickly. According to him, companies around the world are racing to build large-scale AI systems to keep up.

With this move, Marvell becomes more closely aligned with NVIDIA’s growing AI ecosystem, as both companies position themselves to meet increasing demand across enterprise, cloud, and telecom markets.

Also Read: NVIDIA Invests $150 Million in AI Inference Firm Baseten