Qualcomm has announced its return to the data center CPU market, developing custom CPUs that will utilize technology from Nvidia to connect seamlessly with Nvidia's AI chips. This move aims to enhance the performance and energy efficiency of data centers.

Nvidia's AI chips, while dominant, have traditionally been paired with CPUs from Intel and Advanced Micro Devices. Nvidia has also ventured into the CPU market with its "Grace" CPU, based on Arm Holdings technology.

Qualcomm's history in the data center CPU market includes a previous attempt in the 2010s, where they developed an Arm-based CPU and tested it with Meta Platforms. However, this project was curtailed due to cost cuts and legal challenges. After acquiring a team of ex-Apple chip designers in 2021, Qualcomm quietly revived its CPU development efforts and has engaged in discussions with Meta about a data center CPU once again.

Qualcomm recently confirmed a letter of understanding with Saudi Arabian AI firm Humain to develop a custom data center CPU. The future chips will incorporate Nvidia's technology to facilitate rapid communication with Nvidia's GPUs, which are central to its AI chip portfolio.

Cristiano Amon stated that the ability to connect Qualcomm's custom processors to Nvidia's rack-scale architecture advances a shared vision of high-performance, energy-efficient computing for data centers.