微软发布 Maia 300 AI 芯片,宣布明年扩大多吉瓦部署
Today we are announcing our biggest bet yet on our own silicon.
微软首次公布自研 AI 芯片 Maia 300 的具体性能数据与大规模部署计划,标志着云厂商在算力自主化上的关键一步,对行业供应链格局有重要影响。
Today we are announcing our biggest bet yet on our own silicon.
今天,我们宣布了对自家芯片的最大押注。
Starting next calendar year, we will scale our Maia family to multi-gigawatt deployments, making it the primary inference system for our own product and services.
从下一个日历年开始,我们将把我们的 Maia 系列扩展到千兆瓦级部署,使其成为我们自有产品和服务的主要推理系统。
And we’re introducing Maia 300, which delivers 40% more tokens per watt compared to leading GPU-based systems.
同时,我们推出了 Maia 300,与领先的基于 GPU 的系统相比,其每瓦特吞吐量提升了 40%。
We are already deploying Maia 200 across the US, Europe, and Australia, delivering up to 1.7x more tokens per watt across experiences in Teams and Copilot.
我们已经在美、欧和澳大利亚部署了 Maia 200,在 Teams 和 Copilot 的体验中实现了最高 1.7 倍的每瓦特吞吐量提升。
Our goal is to build on the full-stack efficiency gains we’ve already seen, while continuing to bring the latest innovations from across the ecosystem into our fleet.
我们的目标是在我们已经看到的全面栈效率提升的基础上继续发展,同时将来自整个生态系统的最新创新引入我们的设备群。
更进一步:量化金融体系
看懂新闻只是起点——沿量化金融路径,把它变成能交付的工程能力