Chrysolites Tech Academy

Video by freeCodeCamp.org · streamed from YouTube. All rights remain with the creator.

Lesson 2: Inside an AI Data Center

About this lesson

The buildings that train and run large AI models are unlike ordinary data centers, and this lesson explains them piece by piece by following electricity from the utility grid all the way to an individual processor. It is a long course in five parts. The first part deals with power: how electricity arrives at high voltage, how it is stepped down and distributed, and how batteries and backup systems keep server racks supplied when the grid falters. The second part covers cooling, including the mechanics of liquid cooling and the ways heat is finally released to the outside world so that dense clusters of processors do not overheat. Part three turns to networking, looking at the optical links and interconnects that let thousands of chips exchange data fast enough to train a model together and to answer requests afterwards. Part four steps back to the building itself and the engineering, purchasing and construction work behind sites that draw power on the scale of gigawatts. The final part examines the money: how computing capacity is hosted and rented, how long-term capacity contracts work, and how the investment in all this infrastructure is justified. The course suits anyone curious about the physical side of AI, and it assumes no background in electrical or mechanical engineering. When you finish, read Chapter 2 of the course book, which works through the key calculations for power, rack counts, cooling water and running costs, with a power-path diagram and a comparison table of cooling methods.

Install Chrysolites Tech Academy

Android / Chrome / Edge: open the browser menu and choose Install app or Add to Home screen.

iPhone / iPad (Safari): tap the Share button, then Add to Home Screen.