How AI is changing the way the world builds computers

Tech companies are now packing GPUs — which are ideal for running the calculations that power AI — as tightly as possible into specialized computers.

The result is a new kind of supercomputer — a collection of up to 100,000 chips wired together in buildings known as data centres to hammer away at making powerful AI systems.

All this computing power comes at a cost. OpenAI, the maker of ChatGPT, hopes to build about five facilities that would collectively consume more electricity than the roughly 3 million households in Massachusetts.

As technology companies chase the dream of AI, these data centres are popping up across the country and around the globe, forcing tech giants to hunt for the electricity to power them and the water for cooling systems to keep the chips from frying in their own heat.

Live Events

This is the most fundamental change to computing since the early days of the World Wide Web. Just as companies completely rebuilt their computer systems to accommodate the new commercial internet in the 1990s, they are now rebuilding from the bottom up — from tiny components to the way that computers are housed and powered — to accommodate AI.

Big tech companies have constructed computer data centres all over the world for two decades. The centres have been packed with computers to handle the online traffic flooding into the companies' internet services, including search engines, email applications and e-commerce sites.

But those facilities were lightweights compared with what's coming. Back in 2006,