August 2026
Intermediate
368 pages
8h 6m
English
Content preview from AI Powered Digital Twins
Become an O’Reilly member and get unlimited access to this title plus top books and audiobooks from O’Reilly and nearly 200 top publishers, thousands of courses curated by job role, 150+ live events each month,
O’Reilly covers everything we've got, with content to help us build a world-class technology community, upgrade the capabilities and competencies of our teams, and improve overall team performance as well as their engagement.I wanted to learn C and C++, but it didn't click for me until I picked up an O'Reilly book. When I went on the O’Reilly platform, I was astonished to find all the books there, plus live events and sandboxes so you could play around with the technology.I’ve been on the O’Reilly platform for more than eight years. I use a couple of learning platforms, but I'm on O'Reilly more than anybody else. When you're there, you start learning. I'm never disappointed.I'm always learning. So when I got on to O'Reilly, I was like a kid in a candy store. There are playlists. There are answers. There's on-demand training. It's worth its weight in gold, in terms of what it allows me to do.
Chapter 5 Energy Infrastructure for AI at Scale
AI-powered digital twins deliver transformative value only when they power on. AI’s primary constraints are gigawatts (rather than algorithms, talent, or capital). Energy availability determines what’s computationally possible. Every simulation, inference, and agent requires electrons flowing reliably at massive scale. The path forward requires portfolio approaches combining efficiency optimization, energy generation investments, and architectural decisions. Ultimately, we need to calculate the power required for our digital twin architecture.
Power Requirements and Cooling Infrastructure
Traditional data centers consume 3–7 kW per rack with air cooling sufficient for heat dissipation. AI data centers operate fundamentally differently: GPU servers require 70–132 kW per rack—nearly 20 times traditional density—versus 3–7 kW for conventional servers. This density makes air cooling physically impossible—heat flux exceeds what air can dissipate without creating dangerous temperature gradients. Liquid cooling is more appropriate for AI facilities. Direct-to-chip cooling circulates chilled liquid through cold plates mounted on processors, extracting heat at the source before it enters the data center environment. Immersion cooling submerges entire servers in dielectric fluid, eliminating fans and dramatically improving thermal efficiency. These techniques reduce cooling energy from 40% of total facility power consumption in air-cooled ...
Become an O’Reilly member and get unlimited access to this title plus top books and audiobooks from O’Reilly and nearly 200 top publishers, thousands of courses curated by job role, 150+ live events each month,
and much more.
Read now
Unlock full access
More than 5,000 organizations count on O’Reilly
O’Reilly covers everything we've got, with content to help us build a world-class technology community, upgrade the capabilities and competencies of our teams, and improve overall team performance as well as their engagement.Julian F.
I wanted to learn C and C++, but it didn't click for me until I picked up an O'Reilly book. When I went on the O’Reilly platform, I was astonished to find all the books there, plus live events and sandboxes so you could play around with the technology.Addison B.
I’ve been on the O’Reilly platform for more than eight years. I use a couple of learning platforms, but I'm on O'Reilly more than anybody else. When you're there, you start learning. I'm never disappointed.Amir M.
I'm always learning. So when I got on to O'Reilly, I was like a kid in a candy store. There are playlists. There are answers. There's on-demand training. It's worth its weight in gold, in terms of what it allows me to do.