HotChips Take: HBF Limited to Sparse Attention, Efficiency > Cost
BenBajarin · x · 2026-08-24
Analysis from HotChips 2026 suggests HBF is overhyped. While offering high capacity and low cost, its use is limited to areas like Sparse Attention (DeepSeek/ZAI style) or Expert parallelism. The metric of tokens per second per watt is argued to be more critical than cost per GB.
Related event: HBF High-Bandwidth Flash Faces Doubts Over Limited Use Cases(2 posts)→
More from Infra
- engy.ai processes 50B+ tokens in a day as open-model inference demand surges — const_reborn · 2026-10-16
- New inference engine 'e' hits 10,500 tok/s on a single DGX Station — brandon_galang · 2026-10-16
- SGLang ships Day-0 support for FLUX 3 Image Klein: 10 reference images, BF16/FP8/NVFP4 variants — ying11231 · 2026-10-16
- NVIDIA's Nemotron 3.5 Super VL hits CoreWeave: 12B-active multimodal MoE at $0.25/M input tokens — wandb · 2026-10-16
- Chip equipment beats bull case: 2026 WFE seen at $153B, $278B by 2030 — BenBajarin · 2026-10-16
- Steam now launches in 5.23s, beating Epic's 5.47s — and vLLM gets roasted for slow startup — AlpinDale · 2026-10-16