Replacing HBM with Cheap Flash for Model Weights
BenBajarin · x · 2026-07-08
Semiconductor analyst account chipstrat proposes a "high-bandwidth flash" concept: using cheap flash memory instead of expensive HBM to store model weights. Ben Bajarin reshared the discussion. The core issue is the potential impact of this approach on HBM demand, the memory chip market, and the cost structure of AI inference.
Related event: Carmack Suggests Using NAND Flash Over HBM for AI Inference(2 posts)→
More from Infra
- AWS tutorial: fleet-wide edge CV model lifecycle with SageMaker SDK v3, ~$0.07 per run — AWS ML Blog · 2026-10-27
- Qwen3.8-27B hits 40 tok/s on Jetson Thor with INT4 + MTP, accuracy beats FP16 — No-Dragonfly6246 · 2026-10-27
- Celesto launches CI runners nearly 90% cheaper than GitHub Actions, with agent sandboxes — aniketmaurya · 2026-10-27
- Supabase founders on rebuilding the backend for a billion AI-built apps — Latent Space · 2026-10-27
- AW-0 runs at 230ms on a single RTX 5090, 2.4x faster than autoregressive code planning — trevordarrell · 2026-10-27
- Site tracks US city council topics: data centers rank only behind pickleball courts — zakkohane · 2026-10-27