<?xml version="1.0" encoding="UTF-8"?>
<rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom">
  <channel>
    <title>Google Tensor Processing Units (TPUs) — FoundationIndex</title>
    <link>https://foundationindex.ustunozgur.com/s/tpu</link>
    <description>A Tensor Processing Unit is a Google-designed chip that speeds up the matrix operations used to train and run neural networks. Google connects multiple TPUs so they can work on models that exceed the capacity of a single chip.</description>
    <language>en</language>
    <atom:link href="https://foundationindex.ustunozgur.com/s/tpu/feed.xml" rel="self" type="application/rss+xml" />
    <item>
      <title>Google Tensor Processing Units (TPUs) · Milestone · 2026-11-09</title>
      <link>https://foundationindex.ustunozgur.com/s/tpu#tpu-8tx-8i-max-2026</link>
      <guid>https://foundationindex.ustunozgur.com/s/tpu#tpu-8tx-8i-max-2026</guid>
      <pubDate>Mon, 09 Nov 2026 00:00:00 GMT</pubDate>
      <description>&lt;p&gt;Google announced TPU 8tx for confidential inference and TPU 8i Max for long-context inference. The 8i Max pairs two 8i chips on one package, increasing memory bandwidth and SRAM per package.&lt;/p&gt;&lt;p&gt;Sources: &lt;a href=&quot;https://cloud.google.com/blog/topics/ai-infrastructure/tpu-8tx-and-tpu-8i-max&quot;&gt;Our newest TPUs: 8tx and 8i Max, built for the agentic era&lt;/a&gt;&lt;/p&gt;</description>
    </item>
    <item>
      <title>Google Tensor Processing Units (TPUs) · Milestone · 2026-04-22</title>
      <link>https://foundationindex.ustunozgur.com/s/tpu#tpu-eighth-generation-2026</link>
      <guid>https://foundationindex.ustunozgur.com/s/tpu#tpu-eighth-generation-2026</guid>
      <pubDate>Wed, 22 Apr 2026 00:00:00 GMT</pubDate>
      <description>&lt;p&gt;Google described two eighth-generation TPU designs. The 8t emphasizes training throughput across large pods, while the 8i adds on-chip SRAM and specialized collectives for inference.&lt;/p&gt;&lt;p&gt;Sources: &lt;a href=&quot;https://cloud.google.com/blog/products/compute/tpu-8t-and-tpu-8i-technical-deep-dive&quot;&gt;Inside the eighth-generation TPU: An architecture deep dive&lt;/a&gt;&lt;/p&gt;</description>
    </item>
    <item>
      <title>Google Tensor Processing Units (TPUs) · Release · 2025-11-24</title>
      <link>https://foundationindex.ustunozgur.com/s/tpu#tpu-ironwood-2025</link>
      <guid>https://foundationindex.ustunozgur.com/s/tpu#tpu-ironwood-2025</guid>
      <pubDate>Mon, 24 Nov 2025 00:00:00 GMT</pubDate>
      <description>&lt;p&gt;Google Cloud made TPU7x, the first Ironwood variant, generally available. Its higher-memory configuration targets both model training and inference.&lt;/p&gt;&lt;p&gt;Sources: &lt;a href=&quot;https://docs.cloud.google.com/tpu/docs/release-notes#November_24_2025&quot;&gt;Cloud TPU release notes for November 24, 2025&lt;/a&gt;&lt;/p&gt;</description>
    </item>
    <item>
      <title>Google Tensor Processing Units (TPUs) · Paper · 2023-04-04</title>
      <link>https://foundationindex.ustunozgur.com/s/tpu#tpu-v4-paper-2023</link>
      <guid>https://foundationindex.ustunozgur.com/s/tpu#tpu-v4-paper-2023</guid>
      <pubDate>Tue, 04 Apr 2023 00:00:00 GMT</pubDate>
      <description>&lt;p&gt;The TPU v4 paper described optical circuit switches that reconfigure links within a 4,096-chip system. The fabric can accommodate different topologies and route around unavailable components.&lt;/p&gt;&lt;p&gt;Sources: &lt;a href=&quot;https://arxiv.org/abs/2304.01433&quot;&gt;TPU v4: An Optically Reconfigurable Supercomputer for Machine Learning with Hardware Support for Embeddings&lt;/a&gt;&lt;/p&gt;</description>
    </item>
    <item>
      <title>Google Tensor Processing Units (TPUs) · Paper · 2017-04-16</title>
      <link>https://foundationindex.ustunozgur.com/s/tpu#tpu-v1-paper-2017</link>
      <guid>https://foundationindex.ustunozgur.com/s/tpu#tpu-v1-paper-2017</guid>
      <pubDate>Sun, 16 Apr 2017 00:00:00 GMT</pubDate>
      <description>&lt;p&gt;Google described its first data-center TPU, an inference accelerator built around a systolic array for 8-bit matrix multiplication. The study compared it with CPUs and GPUs on deployed neural networks.&lt;/p&gt;&lt;p&gt;Sources: &lt;a href=&quot;https://arxiv.org/abs/1704.04760&quot;&gt;In-Datacenter Performance Analysis of a Tensor Processing Unit&lt;/a&gt;&lt;/p&gt;</description>
    </item>
  </channel>
</rss>
