<?xml version="1.0" encoding="UTF-8"?><rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom"><channel><title>Yestino - The Signal · Maia</title><link>https://yestino.com/entities/maia-697b8b</link><description>Every event involving Maia</description><language>en</language><atom:link href="https://yestino.com/entities/maia-697b8b/feed.xml" rel="self" type="application/rss+xml"/><item><title>Microsoft plans to scale its own Maia AI chips to gigawatt levels by 2027</title><link>https://yestino.com/events/microsoft-plans-to-scale-its-own-maia-ai-chips-to-gigawatt-l-4cc4bb</link><guid isPermaLink="true">https://yestino.com/events/microsoft-plans-to-scale-its-own-maia-ai-chips-to-gigawatt-l-4cc4bb</guid><pubDate>Mon, 12 Oct 2026 15:24:03 GMT</pubDate><description>Microsoft will make its in-house Azure Maia chips the main AI inference system across its products starting in 2027, deploying them at multi-gigawatt scale.
Sources: The Decoder</description></item><item><title>Microsoft is making its homegrown Maia 300 chip the primary engine for AI inference</title><link>https://yestino.com/events/microsoft-is-making-its-homegrown-maia-300-chip-the-primary-ad4907</link><guid isPermaLink="true">https://yestino.com/events/microsoft-is-making-its-homegrown-maia-300-chip-the-primary-ad4907</guid><pubDate>Mon, 12 Oct 2026 13:40:22 GMT</pubDate><description>Starting with the third-generation Maia 300, Microsoft says it will rely on its own silicon to handle incoming AI model requests
Sources: Quartz</description></item></channel></rss>