AMD has reportedly told AIB partners about a minimum 10% hike planned for GPU and memory bundles later this month, following ...

Semiconductor stocks are ripping higher midday Tuesday as a broad risk-on tape lifts the entire chip complex. Intel ...

芯硬件最新获悉，AMD已经通知AIB品牌，将于8月上调显卡GPU供货价格，涨价幅度至少10%以上。内部供应链消息确认，AMD已于上个月敲定GPU+显存套装的供货提价方案，但当时并未立即执行，核心原因是考虑到显卡市场需求疲弱，还处于复苏阶段，终端对价格敏感度高，而且市场仍然存在一定库存等待消化，贸然提价会打击消费信心。另 ...

文｜晓静编辑｜徐青阳7月24日，AMD在Advancing AI 2026大会上发布Helios机架级AI平台和第六代EPYC“Venice”服务器CPU。其中，Helios搭载新一代Instinct MI455X GPU、Venice ...

据TrendForce此前发布的预测，2026年第二季度传统DRAM合约价环比涨幅可达58%至63%。受供应商产能持续向服务器端倾斜、云服务厂商提前锁定产能影响，消费级DDR4、DDR5内存，以及显卡所用的GDDR6、GDDR7显存，均面临供应收缩与 ...

There's no arguing that AMD has come a long way with its graphics cards, but there's still a gap between it an Nvidia that ...

When Nvidia rival AMD uncorks its earnings report later today, it's reasonable to expect a whole lot of bullishness. That comes as no surprise. AMD has been knocking it out of the ...

AMD has answered Nvidia’s Rubin GPU v challenge with the introduction of the Instinct MI455X accelerator, its flagship ...

The stock is rebounding with the whole AI infrastructure trade, and the article highlights NVDA’s valuation discount versus peers (forward P/E ~18.9 vs PHLX ~20.6; AMD/Intel far higher). That discount ...

目前对比仅限于整型工作负载，浮点性能尚未公布，AMD在当前服务器CPU市场的向量化性能方面占据强势地位，AVX-512能力一直是EPYC的卖点。而AI训练和推理工作负载中矩阵运算密集，浮点和向量指令的性能往往比整型吞吐量更具实际参考意义。

快科技3月18日消息，NVIDIA CPU处理器正在从内部自用，转向内外兼备，既自己用，也对外卖。 这一转变就从当下的Vera CPU开始，基于Arm指令集、自研架构、72核心，结合空间多线程、高速互连等创新技术，号称IPC性能提升1.5倍、单线程性能无敌、多线程效率奇高。

过去一年多，大家最熟悉的是“生成式 AI”——你提问，大模型帮你写文章或画图。但现在的行业焦点，正在向两个更深度的方向演进：一个是“AI智能体（Agentic AI）”，也就是让 AI ...