NVIDIA introduces its new superchips, which offer a combination of Grace CPU and Hopper GPU chips. These chips enable the entire AI and data center ecosystem to reap the benefits. The team at NVIDIA aims to provide customers with a comprehensive package that allows them to focus on different types of workloads in the industry. This is achieved by integrating CPU GPU into a single design. The Grace Superchips are available as both GPU CPU and CPU-only variants in a 2-chip package.
However, customers may be limited in some ways, as they are now forced to use either NVIDIA’s Hopper architecture or two CPUs, even if their needs would only require a single chip. However, there are plans by NVIDIA to offer the high-end Grace CPU separately.
NVIDIA still plans to expand its superchip designs in CPU GPU packages with upcoming architectures such as GH200, GB200 and GX200, according to its current GPU roadmap. However, there is a possibility that the company will make a surprise announcement at the upcoming GTC and introduce the Grace CPU or its successor as a standalone server CPU. At a Wells Fargo event, NVIDIA CFO Colette Kress was asked if the company has plans to offer the Grace architecture as a standalone CPU for the server industry. The official response came with a bit of enthusiasm. Here’s what she had to say:
Is it going to be Grace Hopper, GH200, GH300, whatever the subsequent versions might look like? Or is there just a Grace? Is there a market for just an ARM-based CPU from NVIDIA?
Colette Kress (NVIDIA CFO):
There is an opportunity for just a Grace. There is an opportunity for just Grace. New product scenarios that we could see in the data center, you will likely see opportunities for Grace as well.