Qualcomm Adopts Nvidia NVLink Fusion for Data Center Expansion
Qualcomm integrated Nvidia's NVLink Fusion to connect its Oryon CPUs with Blackwell Ultra and Rubin GPUs for large-scale data center AI training and inference.
Qualcomm Inc. adopted Nvidia's NVLink Fusion to integrate its Oryon Arm-based CPUs with Nvidia's Blackwell Ultra and Rubin GPUs. This strategic move allows the company to enter the large-scale data center training and inference market, expanding beyond its current focus on edge AI and the Cloud AI100 Ultra PCIe card.
By utilizing NVLink Fusion, Qualcomm gains a high-speed scale-up interconnect for its future data center CPUs. This allows the company to bypass the lower performance associated with open-source alternatives such as UALink and UltraEthernet.
The arrangement supports Qualcomm's hybrid intelligence strategy, which seeks to combine on-device AI with distributed cloud capabilities. Through this integration, Qualcomm leverages Nvidia's established logic and tooling to scale its hardware capabilities for data center applications.