<?xml version="1.0" encoding="UTF-8"?>
<rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom">
  <channel>
    <title>The Minutes of Dylan Patel on cpu demand</title>
    <link>https://minutesof.com/dylan-patel/on/cpu-demand/</link>
    <description>Everything Dylan Patel has said on cpu demand: 2 verbatim quotes from April 2026, each with a timestamp and a link to the recording it came from.</description>
    <language>en</language>
    <lastBuildDate>Sun, 30 Aug 2026 16:22:22 +0000</lastBuildDate>
    <atom:link href="https://minutesof.com/dylan-patel/on/cpu-demand/feed.xml" rel="self" type="application/rss+xml" />
    <item>
      <title>Patel says CPU-to-GPU ratio has shifted from 1:100 megawatts to much closer parity.</title>
      <link>https://minutesof.com/q/8f0a6563-524e-4327-a1d7-06032c4f32c0/</link>
      <guid isPermaLink="true">https://minutesof.com/q/8f0a6563-524e-4327-a1d7-06032c4f32c0/</guid>
      <description>“So whereas before you&#x27;d have like many GPU servers per CPU server, and so you&#x27;d be like, 100 megawatts of GPUs would be served by even like one megawatt or less of CPUs.” — Daytona</description>
      <pubDate>Tue, 07 Apr 2026 17:12:37 +0000</pubDate>
      <category>cpu demand</category>
    </item>
    <item>
      <title>Patel says CPU-to-GPU ratios have tightened from 1:100 megawatts to much closer ratios due to RL and agentic i</title>
      <link>https://minutesof.com/q/e7053125-a899-4c7e-881d-a55e77a59048/</link>
      <guid isPermaLink="true">https://minutesof.com/q/e7053125-a899-4c7e-881d-a55e77a59048/</guid>
      <description>“Nowadays the ratio is like getting much, much closer, both for RL training and for inference, agentic inference.” — Daytona</description>
      <pubDate>Tue, 07 Apr 2026 17:12:37 +0000</pubDate>
      <category>cpu demand</category>
    </item>
  </channel>
</rss>
