At G20, Big Tech and Global Leaders Can’t Escape AI’s Data Center Problem
At the G20 Innovation Ministerial summit on Wednesday, US Commerce Secretary Howard Lutnick, on stage with Anthropic co-founder and chief compute officer Tom Brown, made a controversial argument: If the world wants to grow AI, we need data centers.
On the surface, this is logical. Data centers are a crucial piece of infrastructure that tech companies need to power their training of new AI models. But the first part of that “if-then” statement is questionable. These massive AI server farms are wildly unpopular, in large part because of serious environmental concerns, energy costs and resource guzzling.
Over the past few years, data center construction projects have been sprouting across the US, with about 3,400 sites already operational and another 2,000 planned, according to Cleanview’s data center tracker tool. The facilities are notorious for demanding massive amounts of electricity to power AI compute, clean water to cool the computer chips...
Copyright of this story solely belongs to cnet.com. To see the full text click HERE
