AI芯片成本2/3被内存吃掉
推荐指数 58.0 NO. 013 · 2026.05.25
发布2026/05/24Score132Comments133
为什么值得看
Epoch AI数据显示，内存已占AI芯片组件成本的63%，较2015年的约20%大幅攀升。这意味着算力扩张的瓶颈正从计算单元转向存储带宽，直接影响大模型训练和推理的成本结构。
媒体预览
编辑判断
内存占比飙升的核心推手是HBM（高带宽内存）的堆叠层数竞赛——从HBM2的8层到HBM3E的12层，单颗容量翻了几倍但良率爬坡极慢。台积电CoWoS封装产能不足，导致HBM成为整个AI供应链最硬的瓶颈。
这对读者的直接影响是：买GPU不能只看算力TFLOPS，内存带宽（GB/s）和容量（GB）才是决定你能跑多大模型、什么batch size的关键指标。如果你在做推理优化，现在就该把重点从算子融合转向KV Cache压缩和量化策略，因为内存成本占比只会继续涨。
创业机会在CXL内存扩展和存算一体架构，英伟达垄断HBM控制器的局面不可能永远持续。
社区反馈
负面 126 条评论
核心争论：内存价格暴涨是周期波动还是结构性拐点，AI需求是否彻底改变了存储成本结构
I bought 96GB of RAM a couple of years ago for ~$250. That same RAM now costs $1200!
It is one of the thing with consumer when they remember they brought it at the absolutely lowest price point when DRAM maker were bleeding money. Those are not normal pricing. Before the pricing collapse in early 2020, 96GB DDR5 would have cost about $450 to $500. And I will need to restate again th
> Crucial are already selling DRAM made by CXMT. Crucial was disestablished this year.