“In our last quarter, I think data center was over 50% of our revenue,” AMD CEO Lisa Su said at the recent Goldman Sachs Communacopia And Technology Conference. Adding, “So we really are a data center first company.” Although this might sound like a shift in focus for AMD, with the generative AI boom in full swing, it’s par for the course for chipmakers feeding the insatiable AI beast.
VIEW GALLERY – 2 IMAGES
Looking at AMD’s most recent earnings, data center revenue, which includes server-specific EPYC CPUs and other GPU hardware, brought in roughly double the revenue of AMD’s client and gaming business. Lisa Su added that AMD’s data center-first approach extends to cloud, edge, and client – as seen in the latest Ryzen AI chips for mobile devices.
It’s a similar story at NVIDIA, which has seen unprecedented growth and expansion in recent years due to its data center and AI software business.
The big takeaway is that the push to develop increasingly more powerful AI hardware for data centers and companies like Google, Meta, Amazon, Microsoft, and Oracle means that new and next-gen GPU and CPU architecture is being designed with this in mind. From there, it will filter down to the client-based products we know, like Ryzen, Radeon, and GeForce RTX on the NVIDIA side.
And we’re already seeing the results of this, with AMD recently announcing that it was combining its RDNA graphics and CDNA compute into a unified UDNA architecture – like NVIDIA with CUDA. This means its graphics technology across Radeon GPUs for desktop and data center GPUs will become the same – excellent news for Radeon fans as a more significant investment in all AMD graphics could lead to a more impressive GPU lineup for desktops, laptops, and handhelds.