<?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>盈亏平衡 on AI 实战派 · 从技术到赚钱</title><link>https://guijiagi.com/tags/%E7%9B%88%E4%BA%8F%E5%B9%B3%E8%A1%A1/</link><description>Recent content in 盈亏平衡 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>Fri, 09 Oct 2026 14:48:00 +0800</lastBuildDate><atom:link href="https://guijiagi.com/tags/%E7%9B%88%E4%BA%8F%E5%B9%B3%E8%A1%A1/index.xml" rel="self" type="application/rss+xml"/><item><title>用 Anthropic 的账倒推：一家前沿模型公司，到底要做到多少营收才能打平</title><link>https://guijiagi.com/posts/2026-10-09-19-frontier-lab-breakeven-revenue/</link><pubDate>Fri, 09 Oct 2026 14:48:00 +0800</pubDate><guid>https://guijiagi.com/posts/2026-10-09-19-frontier-lab-breakeven-revenue/</guid><description>Anthropic 2025 年营收 45.9 亿、营业亏损 80.6 亿、算力支出 73.3 亿。用这三个数字倒推，前沿实验室的盈亏平衡点远比想象中高。</description></item><item><title>用 SemiAnalysis 的 $0.123 算清账：什么时候才值得租一台 GB300 NVL72</title><link>https://guijiagi.com/posts/2026-10-09-02-gb300-break-even-spreadsheet/</link><pubDate>Fri, 09 Oct 2026 11:06:00 +0800</pubDate><guid>https://guijiagi.com/posts/2026-10-09-02-gb300-break-even-spreadsheet/</guid><description>NVIDIA 称 GB300 NVL72 推理成本低至 $0.123/百万 token。本文把这个数字拆成一张自建 vs 云 API 的盈亏平衡表，算清日 Token 量到多少才值得自建。</description></item></channel></rss>