“I believe data center accounted for more than 50% of our revenue in our most recent quarter,” AMD CEO Lisa Su stated at the most recent Goldman Sachs Communacopia and Technology Conference. Saying, “So we really are a data center first company.” With the generative AI boom in full swing, this may sound like a change of direction for AMD, but chipmakers feeding the ravenous AI beast are not unusual in this regard.
AMD’s latest financial results show that revenue from data centers, including server-focused EPYC CPUs and various GPUs, was roughly twice that of its client and gaming sectors. Su further explained that AMD’s data center-first approach also impacts its cloud, edge, and client technologies, such as the new Ryzen AI chips designed for mobile devices.
This shift is mirrored by NVIDIA, which has seen substantial growth and expansion driven by its data center and AI software ventures. The growing focus on developing high-performance AI hardware for major tech firms like Google, Meta, Amazon, Microsoft, and Oracle is influencing the design of next-generation GPUs and CPUs.
Recently, AMD unveiled its unified UDNA architecture, combining CDNA computing and RDNA graphics. This consolidation aims to unify AMD’s graphics technology across its Radeon GPUs for both desktop and data center use, which could enhance the GPU offerings for desktops, laptops, and handheld devices.