PulseAugur
EN
LIVE 20:28:39

NVIDIA Rubin Ultra GPU preview featured 1TB HBM due to supply, not performance

By PulseAugur Editorial · [1 sources] ·

SemiAnalysis discussed NVIDIA's Rubin Ultra GPU, highlighting a preview that featured 1TB of High Bandwidth Memory (HBM) per package, significantly exceeding the 192GB found in currently shipping products. The conversation, featuring Myron Xie and Jordan Nanos, suggested this difference is primarily due to supply constraints rather than performance limitations. Xie explained how a 4-High (4-HI) stacking method increases the number of usable chips harvested per wafer, and noted that the demand for less memory is ironically coming from research labs themselves, anticipating a future where memory shortages ease. AI

IMPACT Provides insight into the memory configurations and supply chain considerations for future high-performance GPUs, relevant for AI hardware planning.

RANK_REASON Discussion of a product preview and technical details, not an official release or announcement.

Read on X — SemiAnalysis →

AI-generated summary · Google Gemini · from 1 sources. How we write summaries →

NVIDIA Rubin Ultra GPU preview featured 1TB HBM due to supply, not performance

How we ranked this

Signal score
2 / 100
Composite score across the factors below. Higher = stronger signal that this story matters right now.
Newsworthiness bucket
Commentary
Discussion of a product preview and technical details, not an official release or announcement.
Source corroboration
Single-source cluster
Only one publisher covered this so far. Single-source stories can still rank when the publisher is high-authority, but they lack cross-source corroboration.
Topics
infra, product
Editorial topic classification. Feeds into how the story surfaces on /topic/<slug> hub pages and into the per-entity coverage mix.
AI-industry relevance
High
Clearly on-topic for AI-industry coverage.
Story freshness
Breaking (< 6h)
Fresh story with cross-source coverage still developing. Ranking may shift as more sources report.

Full methodology in our editorial standards.

COVERAGE [1]

  1. X — SemiAnalysis TIER_1 English(EN) · SemiAnalysis_ ·

    Myron Xie joins @JordanNanos on why NVDA's Rubin Ultra preview had 1TB of HBM per package vs the 192GB that ships.

    Myron Xie joins @JordanNanos on why NVDA's Rubin Ultra preview had 1TB of HBM per package vs the 192GB that ships. It comes down to supply and not performance, why 4-HI more than doubles the cubes you can harvest per wafer, why the loudest calls for less memory are coming from

RELATED ENTITIES

RELATED TOPICS