NVIDIA Unveils NVLink Fusion for Industry to Build Semi-Custom AI Infrastructure With NVIDIA Partner Ecosystem

NVIDIA has launched NVLink Fusion, allowing industries to create semi-custom Artificial Intelligence (AI) infrastructure by leveraging partnerships within the NVIDIA ecosystem. This new silicon aims to bolster scalability and performance for AI workloads, facilitating integration with CPUs from Fujitsu and Qualcomm to enhance data center efficiency.

Key collaborators in this initiative include MediaTek, Marvell, Alchip Technologies, Astera Labs, Synopsys, and Cadence, all of which are set to adopt NVLink Fusion to enable custom silicon that meets the demands of various workloads. The integration capabilities of NVLink Fusion allow Fujitsu and Qualcomm's CPUs to work in tandem with NVIDIA GPUs to establish advanced AI factories.

Jensen Huang, founder and CEO of NVIDIA, remarked, “A tectonic shift is underway: for the first time in decades, data centers must be fundamentally rearchitected — AI is being fused into every computing platform.” This initiative opens doors for partners to build specialized AI infrastructures, addressing evolving industry needs.

The NVLink Fusion technology provides cloud providers the means to expand AI factories efficiently, utilizing NVIDIA’s rack-scale systems and an extensive networking platform that supports significant throughput. Hyperscalers can now utilize the NVIDIA partner ecosystem to integrate these solutions seamlessly within their data center infrastructures.

MediaTek's CEO, Rick Tsai, stated, “By leveraging our world-class ASIC design services and deep expertise in high-speed interconnects, MediaTek is collaborating with NVIDIA to build the next generation of AI infrastructure.” Similarly, other partners, including Marvell, Alchip, Astera Labs, Synopsys, and Cadence, highlighted their commitments to develop custom solutions that maximize the potential of NVLink Fusion.

NVIDIA's NVLink platform aims to provide comprehensive support for AI chip development, ensuring efficient deployment and enabling rapid model training. The initiative reflects a strategy focused on creating robust AI infrastructures that meet future demands.

NVIDIA NVLink Fusion design services are now available through its partner network, reinforcing the collaborative approach to scaling AI capabilities in cloud environments.