<?xml version="1.0" encoding="UTF-8"?>
<rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom" xmlns:dc="http://purl.org/dc/elements/1.1/">
<channel>
<title>ClearCompute blog</title>
<link>https://clearcompute.cc/blog</link>
<atom:link href="https://clearcompute.cc/blog/feed.xml" rel="self" type="application/rss+xml"/>
<description>Practical writing on running AI inference, from ClearCompute.</description>
<language>en-gb</language>
<item>
<title>What an H100 costs per hour in October 2026, and what the price leaves out</title>
<link>https://clearcompute.cc/blog/h100-cloud-prices</link>
<guid isPermaLink="true">https://clearcompute.cc/blog/h100-cloud-prices</guid>
<pubDate>Mon, 05 Oct 2026 09:00:00 GMT</pubDate>
<dc:creator>Ronit Ahuja</dc:creator>
<description>On 5 October 2026 one H100 listed at $2.69 to $12.29 per GPU-hour across nine clouds, with a $1.80 marketplace median. Where the spread comes from, and the costs the hourly rate leaves out.</description>
</item>
<item>
<title>How much GPU memory does a 70B model need? Fitting it is the easy part</title>
<link>https://clearcompute.cc/blog/gpu-memory-70b-model</link>
<guid isPermaLink="true">https://clearcompute.cc/blog/gpu-memory-70b-model</guid>
<pubDate>Mon, 05 Oct 2026 09:00:00 GMT</pubDate>
<dc:creator>Ronit Ahuja</dc:creator>
<description>Llama 3.3 70B needs 141 GB of weights at BF16 and 73 GB at FP8, and every token of a user's context adds 320 KiB of KV cache. The arithmetic for how many users fit on an H100, two H100s, an H200 or a B200.</description>
</item>
<item>
<title>Moving off a closed API: when do your own GPUs pay off?</title>
<link>https://clearcompute.cc/blog/self-hosting-vs-api-cost</link>
<guid isPermaLink="true">https://clearcompute.cc/blog/self-hosting-vs-api-cost</guid>
<pubDate>Mon, 05 Oct 2026 09:00:00 GMT</pubDate>
<dc:creator>Ronit Ahuja</dc:creator>
<description>We priced the same Llama 3.3 70B traffic three ways on 5 October 2026: a closed API, a per-token open-model API and dedicated H100s. Dedicated GPUs beat Sonnet-class prices above about 4% utilisation; for small or spiky traffic on a widely hosted model, the cheapest per-token hosts still cost less.</description>
</item>
</channel>
</rss>
