/
Navigation
Chronicles
Browse all articles
Explore
Semantic exploration
Research
Entity momentum
Nexus
Correlations & relationships
Story Arc
Topic evolution
Drift Map
Semantic trajectory animation
Posts
Analysis & commentary
Pulse API
Tech news intelligence API
Browse
Entities
Companies, people, products, technologies
Domains
Browse by publication source
Handles
Browse by social media handle
Detection
Concept Search
Semantic similarity search
High Impact Stories
Top coverage by position
Sentiment Analysis
Positive/negative coverage
Anomaly Detection
Unusual coverage patterns
Analysis
Rivalry Report
Compare two entities head-to-head
Semantic Pivots
Narrative discontinuities
Crisis Response
Event recovery patterns
Connected
Search: /
Command: ⌘K
Embeddings: large
TEXXR

Chronicles

The story behind the story

← → days · ↑ ↓ browse · Enter similar · o open

Nvidia unveils Grace, a high-performance Arm-based server CPU for large-scale neural network workloads, expected to become available in Nvidia products in 2023

Kicking off another busy Spring GPU Technology Conference for NVIDIA, this morning the graphics and accelerator designer …

AnandTech Ryan Smith

Context & Ripple Effects

Grace marks Nvidia's move from supplying accelerators alongside other vendors' server CPUs toward offering its own Arm-based CPU for neural-network systems. The initial product was later expanded into a 144-core Grace CPU Superchip, showing that the CPU was a product line rather than a one-off component.

The arc then shifted from a standalone CPU to tightly coupled compute packages: Grace Hopper combined Grace with a GPU, and Blackwell's GB200 again paired GPUs with Grace. That makes this announcement the starting point for Nvidia's CPU-GPU system strategy.

First-order effects

  • Nvidia adds an Arm-based server-CPU option to its data-center portfolio, giving customers building large neural-network workloads a CPU designed around Nvidia's own products.
  • Server buyers planning Nvidia-based AI systems gain a new CPU road map, while Nvidia takes on responsibility for the host processor as well as the accelerator.

Second-order effects

  • Nvidia's subsequent Grace Superchip and Grace Hopper products turn CPU selection into part of a broader Nvidia platform decision, increasing the value of validated CPU-GPU combinations for buyers.
  • Arm's server-core roadmap gains a prominent deployment path: Neoverse V2 was slated to power the upcoming Grace CPU, linking Arm's cloud and HPC ambitions to Nvidia's system products.

Third-order effects

  • If the Grace-to-GH200-to-GB200 progression holds, AI infrastructure will be organized increasingly around heterogeneous superchips rather than separately sourced CPUs and GPUs.
  • That integration gives Nvidia more control over the compute stack, while narrowing the portion of an AI server that customers can choose independently.

The trend: Nvidia is moving from GPU supplier to provider of integrated Arm CPU-GPU systems for AI and high-performance computing.

Discussion

  • @fox0x01 Azeria on x
    NVIDIA announced its new server SoC Grace, which will be based on a future generation of Arm's Neoverse cores. The current Neoverse N1 is based on Armv8.2. I wonder if the N2 will be based on the new Armv9 architecture? https://twitter.com/...
  • @nvidiagtc Nvidia Gtc on x
    Meet the NVIDIA Grace #CPU, leveraging the flexibility of @Arm's data center architecture and designed from the ground up to accelerate the largest HPC and AI workloads. #GTC21 https://nvidianews.nvidia.com/ ... https://twitter.com/...
  • @stevemoser Steve Moser on x
    I can't wait for the Nvidia powered Dell laptop vs M1 MacBook comparison. Nvidia could make a really interesting APU. https://twitter.com/...
  • @dinabass Dina Bass on x
    Another key vendor trying to chip away at Intel's server chip hegemony — one of the key things to watch for the next few years is how this goes (along with efforts from Amazon & others) and whether Pat Gelsinger can beat back the challenge: https://www.bloomberg.com/... by @ianmk…
  • @dakami Dan Kaminsky on x
    Nvidia Grace provides 200GB/sec between CPU and GPU, with complete cache coherency. That's...gonna completely change how a lot of code is written. I mean, in 2023, but still. Hot damn.