返回精选
AI 精选动态 智能评分 65

CPU 市场复苏:四大驱动力

来源: twitter关注列表
作者: SemiAnalysis (@SemiAnalysis_)
发布于: 2026-06-25
收录于: 2026-06-25
AI 推荐理由
与主流认知相反,CPU 因 AI 工作负载扩展成为新瓶颈,建议阅读原文获取详细数据和供应链影响分析。
核心解读
SemiAnalysis 分析指出,CPU 需求因强化学习、代理模型、上下文内存和 RAG 等四大驱动力同步增长,导致 AI 资本支出模型中 CPU 线偏差近一个数量级;供应方面,AMD 计划 2026 年推出 Venice/Verano(256 核,TSMC N2),Intel Diamond Rapids/Coral Rapids(Intel 18A-P,但失去 SMT),Arm 转向全芯片销售 Phoenix,Meta 为首个客户。
全文
Our CPUs Are Back piece is the work weve been pointing people to whenever they ask why server CPU revenue, which was supposed to be dead, is suddenly behaving like an early-cycle category again. (1/4)🧵 https://t.co/9bgyBTi7NE ![photo](https://pbs.twimg.com/media/HLrLbEKXQAAFE3Y.jpg) SemiAnalysis (@SemiAnalysis_): On the supply side, we map the 2026 roadmap from AMD Venice and Verano (256 cores, TSMC N2) to Diamond Rapids and Coral Rapids (Intel 18A-P, but losing SMT) to Arm pivoting from licensing IP to selling Phoenix as a full chip with Meta as lead customer. The punchline, and where the article ends, is that the CPU line in most AI capex models is off by close to an order of magnitude, and the operators who own silicon on both sides of the cluster (AWS, Azure, Google) are capturing margin twice. (4/4) https://t.co/nP6wILA1VL SemiAnalysis (@SemiAnalysis_): What we walk through in the article is how four demand drivers are stacking at the same time: reinforcement learning, which needs CPU clusters running compilation and verification to keep accelerators busy; agentic models, which call tools and hit databases far more aggressively than any human session did; context memory, where managing petabytes of GPU intermediate state is a real CPU line item; and vibe coding plus RAG, which spread CPU load across many cheap user sessions. (3/4)
#技术#分析#AI