<?xml version="1.0" encoding="utf-8" standalone="yes"?><rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom"><channel><title>Hetzner on PixelUnion - Free your photos from American tech platforms</title><link>https://pixelunion.eu/tags/hetzner/</link><description>Recent content in Hetzner on PixelUnion - Free your photos from American tech platforms</description><generator>Hugo</generator><language>en</language><lastBuildDate>Tue, 16 Jun 2026 21:33:15 +0200</lastBuildDate><atom:link href="https://pixelunion.eu/tags/hetzner/index.xml" rel="self" type="application/rss+xml"/><item><title>Why Cloud Costs Are Exploding</title><link>https://pixelunion.eu/blog/2026/06/cloud-cost/</link><pubDate>Tue, 16 Jun 2026 00:00:00 +0000</pubDate><guid>https://pixelunion.eu/blog/2026/06/cloud-cost/</guid><description>&lt;p>The rapid rise of AI isn&amp;rsquo;t just changing how we work. It&amp;rsquo;s quietly driving up the cost of the infrastructure that powers everything we do online. In 2026, cloud and hardware costs have risen sharply across the board, and the main culprit is a global shortage of the components that make modern computing possible.&lt;/p>
&lt;h2 id="the-ai-industrys-appetite-for-memory">The AI industry&amp;rsquo;s appetite for memory&lt;/h2>
&lt;p>The root of the problem lies in how the semiconductor market has shifted. The three major memory manufacturers, Samsung, SK Hynix and Micron, have redirected a significant share of their production capacity toward High Bandwidth Memory (HBM). This is the specific type of memory that powers AI servers like those used by Nvidia. The side effect is that far less capacity is available for the standard components used in regular servers and PCs.&lt;/p></description></item></channel></rss>