<?xml version="1.0" encoding="utf-8" standalone="yes"?><rss version="2.0" xmlns:atom="http://www.w3.org/2005/Atom"><channel><title>Gcp on Dinesh Mannam</title><link>https://dineshmannam.com/tags/gcp/</link><description>Recent content in Gcp on Dinesh Mannam</description><generator>Hugo</generator><language>en-us</language><managingEditor>dinesh@dineshmannam.com (Dinesh Mannam)</managingEditor><webMaster>dinesh@dineshmannam.com (Dinesh Mannam)</webMaster><lastBuildDate>Sat, 12 Sep 2026 00:00:00 +0000</lastBuildDate><atom:link href="https://dineshmannam.com/tags/gcp/index.xml" rel="self" type="application/rss+xml"/><item><title>What an AI agent actually costs on Google Cloud</title><link>https://dineshmannam.com/engineering/what-an-ai-agent-actually-costs-on-google-cloud/</link><pubDate>Sat, 12 Sep 2026 00:00:00 +0000</pubDate><author>dinesh@dineshmannam.com (Dinesh Mannam)</author><guid>https://dineshmannam.com/engineering/what-an-ai-agent-actually-costs-on-google-cloud/</guid><description>&lt;p&gt;I deployed an ADK agent on Google Cloud, ran it for three days, and pulled the cost from the billing export instead of estimating it. I built the experiment around three levers, two of them outside the model. The model was 87.7% of the bill, and 77.4% of everything it wrote never appeared in a response. The thing that dominated the bill was not on my list.&lt;/p&gt;
&lt;p&gt;The video makes the argument. This is the version you can check. Every number here comes from a query in the repo.&lt;/p&gt;</description></item></channel></rss>