<?xml version="1.0" encoding="utf-8" standalone="yes"?><rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom" xmlns:content="http://purl.org/rss/1.0/modules/content/"><channel><title>吉瓦 on AI 实战派 · 从技术到赚钱</title><link>https://guijiagi.com/tags/%E5%90%89%E7%93%A6/</link><description>Recent content in 吉瓦 on AI 实战派 · 从技术到赚钱</description><generator>Hugo</generator><language>zh-cn</language><copyright>本站内容采用 CC BY-NC-SA 4.0 国际许可协议授权</copyright><lastBuildDate>Mon, 14 Sep 2026 00:00:00 +0000</lastBuildDate><atom:link href="https://guijiagi.com/tags/%E5%90%89%E7%93%A6/index.xml" rel="self" type="application/rss+xml"/><item><title>AI基建的电力约束：吉瓦级数据中心的能源挑战</title><link>https://guijiagi.com/posts/ai-infrastructure-power-constraints-gigawatt/</link><pubDate>Mon, 14 Sep 2026 00:00:00 +0000</pubDate><guid>https://guijiagi.com/posts/ai-infrastructure-power-constraints-gigawatt/</guid><description>Anthropic锁定至少14.8吉瓦计算能力（超1000万户家庭用电）；1吉瓦约对应75万户美国家庭用电；Nscale西弗吉尼亚园区1.35吉瓦；谷歌芬兰购核电50%发电量。当推理算力占比超70%，数据中心能耗已从技术问题升级为两党政治议题。</description></item></channel></rss>