Today for AI
HOT RADAR
computeHEAT 11.3°

AI CURATED HOT STORY · 10/26/2026

NVIDIA Lowers Rubin GPU HBM and NVLink Bandwidth Specs

1 reports archived1 independent sourcesupdated 10/26/2026, 11:26:52
Synthesis & Latest Updates
1 Sources Cross-Validated

NVIDIA's latest documents reveal that the Rubin GPU system's HBM4 memory bandwidth has been reduced from 44TB/s to 38.5TB/s, and NVLink bandwidth from 7.2TB/s to 6TB/s. This adjustment aims to relax hardware requirements to improve HBM4 yield and final production volume, while scale-out network bandwidth sees a slight increase.

LATEST/NVIDIA's latest documents reveal that the Rubin GPU system's HBM4 memory bandwidth has been reduced from 44TB/s to 38.5TB/s, and NVLink bandwidth from 7.2TB/s to 6TB/s. This adjustment aims to relax hardware requirements to improve HBM4 yield and final production volume, while scale-out network bandwidth sees a slight increase.

Hourly heat curve

3 fully observed hours
Now
11.3
Peak
11.812:00
24h change
–

The line compares only sources observed throughout; gap hours are interpolated to keep the trend continuous.

Coverage timeline

Total 1 reports · Latest first
  1. IT之家 · 智能时代T2·78 pts

    NVIDIA Lowers Rubin GPU HBM and NVLink Bandwidth Specs

    • The Vera Rubin superchip's HBM4 bandwidth is reduced from 44TB/s to 38.5TB/s, and NVLink bandwidth from 7.2TB/s to 6TB/s.
    • Lowering bandwidth specs helps improve HBM4 memory compatibility and yield, ensuring large-scale mass production capabilities.
    • Non-superchip solutions (e.g., HGX NVL8) see simultaneous reductions in VRAM capacity and bandwidth, but bidirectional network bandwidth increases to 1.8TB/s.