<!-- mobian-agent-page publisher="dailydev" canonical="https://daily.dev/posts/your-datacenter-s-power-architecture-called-it-s-not-happy-vtkyrqhdc" -->

---
title: Your datacenter&#x27;s power architecture called. It&#x27;s not happy
description: AI workloads like the Nvidia GB200 NVL72 require 120–192 kW per rack, far exceeding what traditional 48V datacenter power architectures can efficiently handle....
canonical: https://daily.dev/posts/your-datacenter-s-power-architecture-called-it-s-not-happy-vtkyrqhdc
twitter:card: summary_large_image
twitter:site: @dailydotdev
og:type: website
og:site_name: daily.dev
og:title: Your datacenter&#x27;s power architecture called. It&#x27;s not happy | daily.dev
og:description: AI workloads like the Nvidia GB200 NVL72 require 120–192 kW per rack, far exceeding what traditional 48V datacenter power architectures can efficiently handle....
og:url: https://daily.dev/posts/your-datacenter-s-power-architecture-called-it-s-not-happy-vtkyrqhdc
og:image: https://api.daily.dev/og/posts/VtKyrQHDc.png
og:image:alt: Your datacenter&#x27;s power architecture called. It&#x27;s not happy
og:image:width: 1200
og:image:height: 630
og:locale: en
---

> ## Documentation Index
> Fetch the complete documentation index at: https://daily.dev/llms.txt
> Use this file to discover all available pages before exploring further.

# Your datacenter's power architecture called. It's not happy

**[The Register](https://daily.dev/sources/theregister)** · 8 min read · 0 upvotes · 0 comments

## Summary

AI workloads like the Nvidia GB200 NVL72 require 120–192 kW per rack, far exceeding what traditional 48V datacenter power architectures can efficiently handle. The core problem is Joule resistive loss: at 48V, delivering hundreds of kilowatts requires thousands of amperes, causing massive copper busbars, heat buildup, and efficiency losses. Moving to 800V DC distribution reduces current by roughly 16x, cutting resistive losses by orders of magnitude and enabling up to 45% copper reduction. The industry is transitioning toward HVDC architectures where AC is converted to 800VDC at the facility perimeter and distributed throughout. New power electronics using GaN and SiC semiconductors handle the step-down to chip-level voltages. Transient load spikes from synchronized GPU operations add further complexity, requiring integrated energy storage. Companies like Eaton, Vertiv, and Delta are already developing 800V-compatible hardware as the industry moves beyond ORv3 limits.

## Full article

daily.dev links to this article rather than hosting it. Read it at the original source: <https://go.theregister.com/feed/www.theregister.com/2026/03/11/your_datacenters_power_architecture_called/>

## Similar posts on daily.dev

- [Data Center DC Embraces 800V Power Shift](https://daily.dev/posts/data-center-dc-embraces-800v-power-shift-brxpg9hrh) · Hacker News · 0 upvotes · 0 comments
- [Building the 800 VDC Ecosystem for Efficient, Scalable AI Factories](https://daily.dev/posts/building-the-800-vdc-ecosystem-for-efficient-scalable-ai-factories-np8vi77vc) · NVIDIA Developer · 0 upvotes · 0 comments
- [How 800 VDC for Data Centers Tames AI Power Swings](https://daily.dev/posts/how-800-vdc-for-data-centers-tames-ai-power-swings-zsltzvtfk) · IEEE Spectrum · 0 upvotes · 0 comments
- [DC In The Data Center For A More Efficient Future](https://daily.dev/posts/dc-in-the-data-center-for-a-more-efficient-future-tkghfbdat) · Hackaday · 0 upvotes · 0 comments
- [Why Scaling AI Compute Performance Requires a New Power Architecture](https://daily.dev/posts/why-scaling-ai-compute-performance-requires-a-new-power-architecture-rxa6sxhuf) · NVIDIA · 0 upvotes · 0 comments

---

Tags: [#nvidia](https://daily.dev/tags/nvidia)

[View this post on daily.dev](https://daily.dev/posts/your-datacenter-s-power-architecture-called-it-s-not-happy-vtkyrqhdc)

```json
{"@context":"https://schema.org","@graph":[{"@type":"Organization","@id":"https://daily.dev/#organization","name":"daily.dev","url":"https://daily.dev","logo":{"@type":"ImageObject","url":"https://daily.dev/apple-touch-icon.png","width":180,"height":180},"sameAs":["https://twitter.com/dailydotdev","https://github.com/dailydotdev","https://www.linkedin.com/company/daily-dev-ltd"]},{"@type":"WebSite","@id":"https://daily.dev/#website","url":"https://daily.dev","name":"daily.dev","publisher":{"@id":"https://daily.dev/#organization"},"potentialAction":{"@type":"SearchAction","target":{"@type":"EntryPoint","urlTemplate":"https://daily.dev/search?q={search_term_string}"},"query-input":"required name=search_term_string"}}]}
{"@context":"https://schema.org","@type":"TechArticle","headline":"Your datacenter's power architecture called. It's not happy","url":"https://daily.dev/posts/your-datacenter-s-power-architecture-called-it-s-not-happy-vtkyrqhdc","mainEntityOfPage":{"@type":"WebPage","@id":"https://daily.dev/posts/your-datacenter-s-power-architecture-called-it-s-not-happy-vtkyrqhdc"},"datePublished":"2026-03-11T11:05:38.094Z","dateModified":"2026-03-11T11:06:14.047Z","description":"AI workloads like the Nvidia GB200 NVL72 require 120–192 kW per rack, far exceeding what traditional 48V datacenter power architectures can efficiently handle....","image":"https://media.daily.dev/image/upload/f_auto,q_auto/v1/posts/e6b7e0af162dd15d6d0ff4c3ed451ab3?_a=AQAEuop","thumbnailUrl":"https://media.daily.dev/image/upload/f_auto,q_auto/v1/posts/e6b7e0af162dd15d6d0ff4c3ed451ab3?_a=AQAEuop","isAccessibleForFree":true,"articleSection":"The Register","inLanguage":"en","publisher":{"@type":"Organization","name":"daily.dev","url":"https://daily.dev","logo":{"@type":"ImageObject","url":"https://daily.dev/apple-touch-icon.png","width":180,"height":180}},"author":{"@type":"Organization","name":"The Register","logo":"https://media.daily.dev/image/upload/t_logo,f_auto/v1/logos/66aa2113fdad463992ffcbf0e8963fda","url":"https://daily.dev/sources/theregister"},"commentCount":0,"discussionUrl":"https://daily.dev/posts/your-datacenter-s-power-architecture-called-it-s-not-happy-vtkyrqhdc","interactionStatistic":[{"@type":"InteractionCounter","interactionType":{"@type":"LikeAction"},"userInteractionCount":0},{"@type":"InteractionCounter","interactionType":{"@type":"CommentAction"},"userInteractionCount":0}],"keywords":"nvidia","timeRequired":"PT8M"}
{"@context":"https://schema.org","@type":"BreadcrumbList","itemListElement":[{"@type":"ListItem","position":1,"name":"Home","item":"https://daily.dev"},{"@type":"ListItem","position":2,"name":"The Register","item":"https://daily.dev/sources/theregister"},{"@type":"ListItem","position":3,"name":"Your datacenter's power architecture called. It's not happy"}]}
```

