The $3 trillion AI datacenter spending spree is only part of the story. A “further infrastructure cost” is looming, and it’s a big one: $720 billion. According to Goldman Sachs, that’s the amount of “grid spending” needed to meet the massive energy demands of this new AI infrastructure.
The scale of the AI build-out is “rapid” and power-hungry. Global datacenter capacity, currently at 59GW, is expected to double by 2030. This year alone, work is expected to start on 10GW of new capacity—a power draw “representing roughly a third of the UK’s power demand.”
This “exponential demand” for AI is creating mega-projects like the $500bn “Stargate” venture and Microsoft’s “world’s most powerful AI datacentre.” These facilities, along with the “incredible” $750bn spend from “hyperscalers,” are all contributing to the massive new load on the world’s power grids.
The $720bn grid cost is a hard, physical-world constraint on the digital boom. It highlights the fact that the AI revolution is not just about code; it’s about concrete, steel, and massive amounts of electricity.
While financial analysts debate the “boom or bubble” question over the $3tn spend, the $720bn grid bill is a non-negotiable cost. The success of the AI boom depends just as much on upgrading our global energy infrastructure as it does on building the datacenters themselves.