Network connectivity per GPU in Vera Rubin deployment
Company: CoreWeave
The claim, verbatim
Each Rubin GPU is served by two ConnectX-9 SuperNICs, delivering up to 1.6 Tb/s of backend network connectivity per GPU
Source (primary)
What It Takes to Bring Up a Multi-Rack NVIDIA Vera Rubin NVL72 Cluster - CoreWeave (-, news_article)
View cached copy (2026-09-19)Live source ↗
Quote: “Each Rubin GPU is served by two ConnectX-9 SuperNICs, delivering up to 1.6 Tb/s of backend network connectivity per GPU”
How we checked this
Checked on September 24, 2026. The cited source supports every part of this claim.
The post describes CoreWeave's Vera Rubin RoCE fabric as giving each Rubin GPU two ConnectX-9 SuperNICs and up to 1.6 Tb/s of backend connectivity.
Confirmed in the source:
- In CoreWeave's Vera Rubin network design, each Rubin GPU is served by two ConnectX-9 SuperNICs
- This delivers up to 1.6 Tb/s of backend network connectivity per GPU
What we did: Read our cached copy of the publisher (https://www.coreweave.com/blog/what-it-takes-to-bring-up-a-multi-rack-nvidia-vera-rubin-nvl72-cluster) in full (17,292 characters, retrieved September 19, 2026) and checked each assertion in the claim against it.
Additional evidence
confirms What It Takes to Bring Up a Multi-Rack NVIDIA Vera Rubin NVL72 Cluster - CoreWeave
Quote: “Each Rubin GPU is served by two ConnectX-9 SuperNICs, delivering up to 1.6 Tb/s of backend network connectivity per GPU”