telexed ~ c / a7fcd2e2-ab3radar:40 · otherLIVE
← back
NO.
#a7fcd2e2
Topic
OTHER
Source
GeekNews
Published
2026-05-24 23:00:49
Importance
★ 4/10 — radar 40
Memory Now Takes Nearly Two-Thirds of AI Chip Component Cost
FIG-0071:1

Memory Now Takes Nearly Two-Thirds of AI Chip Component Cost

AI accelerator economics are shifting from logic toward memory supply. HBM is becoming the cost anchor, so cheaper inference will depend as much on memory capacity as model-side optimization.

[ KEY POINTS ]
  1. HBM rose from 52% of weighted average component cost in Q1 2024 to 63% in Q4 2025 across Nvidia, AMD, Google, and Amazon AI chips.
  2. Logic die share stayed near 13%, so the cost pressure is not coming from compute silicon itself.
  3. Advanced packaging fell from 19% to 15%, and auxiliary parts from 15% to 9%; memory is absorbing the budget share.
  4. Cloud GPU and API pricing will remain sensitive to memory supply. Smaller models and cache-efficient inference keep gaining leverage.
Originalnews.hada.io/topic?id=29837Read original →

// related