NVIDIA has hinted that they might offer a stand-alone version of its Arm-based Grace CPU for the server segment if there's an opportunity.
NVIDIA's Arm-Based Grace CPUs Might Come As Standalone Variants For Data Center Servers
For a quick recap, NVIDIA offers its Superchips with a combination of Grace CPU & Hopper GPU chips to leverage the whole AI and data center ecosystem. Team Green wants to offer its client a whole package to focus on different types of workloads in the industry, by integrating CPU+GPU into a single design. The Grace Superchips come in GPU+CPU and CPU-only offerings in a 2-chip package.
However, this limits customers to an extent, since they are now sort of "bounded" into using NVIDIA's Hopper architecture or two CPUs when their requirement might be just a single chip. However, NVIDIA does have plans to offer its high-end Grace CPU separately.
NVIDIA's current GPU roadmap still shows that the company wants to extend its Superchip design in CPU+GPU packages with upcoming architectures such as GH200, GB200, and GX200. However, we might get a surprise announcement at the upcoming GTC that the company is bringing Grace CPU or its successor to the standalone server segment.
Speaking at the Wells Fargo event, NVIDIA's CFO Colette Kress was asked whether the company has plans to offer its Grace architecture in a standalone CPU for the server industry, and the official did reply with a bit of enthusiasm. Here is what she had to say:
Question:
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.
While NVIDIA's CFO hasn't given an official confirmation, it seems like the company has plans to offer its Grace CPUs separately, judging by how the industry evolves going ahead. In terms of what the Grace CPU offers, it comes with a total of 144 (72 Arm Neoverse V2 per chip) cores, supports up to 960 GB of LPDDR5X memory with up to 1 TB/s of raw bandwidth, and has a combined power draw of 500W. Additional specs include 117 MB of L3 cache, and 58 Gen5 lanes, all while using the TSMC 4N process node.
Offering Grace CPUs in a standalone package could be a game-changer for Team Green, given that their AI products have already segmented deep into the markets. Moreover, it would mean more competition for x86 solutions from companies like Intel and NVIDIA, since official benchmarks have revealed that Grace CPU offers competitive efficiency and performance versus the industry's offerings. Servers aren't the only market that NVIDIA plans to enter with standalone CPUs as the company is expected to enter the PC consumer market with Arm-based solutions by 2025.
About the author: Muhammad Zuhair is a hardware and technology reporter for Wccftech, specializing in the semiconductor industry and the complex interplay between technology, manufacturing, and geopolitics. His coverage focuses on the corporate strategies and technological roadmaps of industry giants like TSMC, NVIDIA, Samsung, and Intel. Zuhair's expertise lies in deconstructing complex topics such as fabrication nodes (e.g., 2nm process), the economic impact of policies like the CHIPS Act, and the strategic development of AI infrastructure from NVIDIA, AMD and Intel.
Follow Wccftech on Google to get more of our news coverage in your feeds.
