# NVIDIA Cuts HBM4 Speed and NVLink Bandwidth Targets for "Vera Rubin" — retained record

> Retained reddit record from reddit:AMD_Stock.

Canonical: https://highsignal.app/data/records/e344e91e549ebaba

[← sources](https://highsignal.app/data)


source: [reddit](https://highsignal.app/data/reddit) (reddit:AMD_Stock)  published 2026-10-26 08:19 ingested 2026-10-27 02:38

[original source ↗](https://www.reddit.com/r/AMD_Stock/comments/1xfqz8a/nvidia_cuts_hbm4_speed_and_nvlink_bandwidth/)

## retained text



```
**Now, NVIDIA's solution is inferior to AMD's rack-scale Helios system in both memory capacity and memory bandwidth. AMD's Instinct MI455X offers significantly more hig**h-**bandwidth memory than NVIDIA's revised "Vera Rubin," with 432 GB of HBM4 per GPU at 23.3 TB/s compared to Rubin's 288 GB at 19.2** TB/s. At the 72-GPU rack level, Helios aggregates roughly 31 TB of HBM4 with about 1.67 PB/s of total memory bandwidth, compared to approximately 20.7 TB and 1.4 PB/s on the Vera Rubin NVL72. The gap extends to scale-up interconnects as well. AMD's UALoE delivers 3.6 TB/s bidirectional per GPU (260 TB/s aggregate), while NVIDIA's NVLink has been reduced to 3 TB/s per GPU (216 TB/s aggregate). These differences give Helios a clear advantage in both on-package memory capacity and system-level bandwidth.
```



document fetched 2026-10-27 00:36 · canonical https://reddit.com/r/AMD_Stock/comments/1xfqz8a/nvidia_cuts_hbm4_speed_and_nvlink_bandwidth
