<?xml version="1.0" encoding="UTF-8"?>
<rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom">
  <channel>
    <title>Custom AI Accelerators — FoundationIndex</title>
    <link>https://foundationindex.ustunozgur.com/s/ai-accelerators</link>
    <description>Custom AI accelerators are chips and systems built around common neural-network operations. They include Google TPUs, AWS Inferentia and Trainium, and wafer-scale designs from Cerebras.</description>
    <language>en</language>
    <atom:link href="https://foundationindex.ustunozgur.com/s/ai-accelerators/feed.xml" rel="self" type="application/rss+xml" />
    <item>
      <title>Custom AI Accelerators · Release · 2025-12-02</title>
      <link>https://foundationindex.ustunozgur.com/s/ai-accelerators#ai-accelerators-trainium3-2025</link>
      <guid>https://foundationindex.ustunozgur.com/s/ai-accelerators#ai-accelerators-trainium3-2025</guid>
      <pubDate>Tue, 02 Dec 2025 00:00:00 GMT</pubDate>
      <description>&lt;p&gt;AWS launched Trn3 UltraServers with up to 144 Trainium3 chips connected by NeuronLink. The scale-up system lets training workloads distribute model state across a shared high-bandwidth interconnect.&lt;/p&gt;&lt;p&gt;Sources: &lt;a href=&quot;https://aws.amazon.com/about-aws/whats-new/2025/12/amazon-ec2-trn3-ultraservers/&quot;&gt;Announcing Amazon EC2 Trn3 UltraServers&lt;/a&gt;&lt;/p&gt;</description>
    </item>
    <item>
      <title>Custom AI Accelerators · Release · 2024-08-27</title>
      <link>https://foundationindex.ustunozgur.com/s/ai-accelerators#ai-accelerators-cerebras-inference-2024</link>
      <guid>https://foundationindex.ustunozgur.com/s/ai-accelerators#ai-accelerators-cerebras-inference-2024</guid>
      <pubDate>Tue, 27 Aug 2024 00:00:00 GMT</pubDate>
      <description>&lt;p&gt;Cerebras opened an inference API serving 8B and 70B Llama 3.1 models in 16-bit precision on its third-generation Wafer Scale Engine. This made wafer-scale inference available as a hosted API.&lt;/p&gt;&lt;p&gt;Sources: &lt;a href=&quot;https://www.cerebras.ai/blog/introducing-cerebras-inference-ai-at-instant-speed&quot;&gt;Introducing Cerebras Inference&lt;/a&gt;&lt;/p&gt;</description>
    </item>
    <item>
      <title>Custom AI Accelerators · Release · 2019-12-03</title>
      <link>https://foundationindex.ustunozgur.com/s/ai-accelerators#ai-accelerators-inferentia-2019</link>
      <guid>https://foundationindex.ustunozgur.com/s/ai-accelerators#ai-accelerators-inferentia-2019</guid>
      <pubDate>Tue, 03 Dec 2019 00:00:00 GMT</pubDate>
      <description>&lt;p&gt;AWS launched EC2 Inf1 instances powered by Inferentia. Customers could compile models with the Neuron SDK and rent inference-specific instances as an alternative to GPU instances.&lt;/p&gt;&lt;p&gt;Sources: &lt;a href=&quot;https://aws.amazon.com/blogs/aws/amazon-ec2-update-inf1-instances-with-aws-inferentia-chips-for-high-performance-cost-effective-inferencing/&quot;&gt;Amazon EC2 Inf1 Instances with AWS Inferentia Chips&lt;/a&gt;&lt;/p&gt;</description>
    </item>
    <item>
      <title>Custom AI Accelerators · Paper · 2017-04-16</title>
      <link>https://foundationindex.ustunozgur.com/s/ai-accelerators#ai-accelerators-tpu-paper-2017</link>
      <guid>https://foundationindex.ustunozgur.com/s/ai-accelerators#ai-accelerators-tpu-paper-2017</guid>
      <pubDate>Sun, 16 Apr 2017 00:00:00 GMT</pubDate>
      <description>&lt;p&gt;Google documented a data-center accelerator built for neural-network inference. Its matrix unit illustrates how a chip tuned for a particular class of computation can spend more die area on arithmetic.&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>
