AI Data Center Spending Hits $736 Among Tech Giants
The race among the world’s AI tech leaders has intensified over the last year. The five largest builders of these centers (known as hyperscalers) will spend $736 billion this year and next. They barely have enough cash on their balance sheets to cover this sum. Even with a substantial annual cash flow, they may still need to borrow.
Data centers took up 1% to 2% of the world’s electricity. Goldman expects this to rise to 3% to 4% by the end of the decade. In the US, which means it will have doubled over the period.
“Our analysts estimate current demand to be approximately 62 gigawatts (GW), comprised of cloud workloads (58%), traditional workloads (29%), and AI workloads (13%),” Goldman Sachs reports that the mix will change considerably in the next two years. In 2027, AI’s share of this market is expected to grow to 28% and continue rising.
Where Does The Money Come From?
The analysis shows that electricity demand in AI data centers will increase by 165% between 2023 and 2030.
The growth recreates a significant emissions problem. As the need for energy rises, as much as 60% of the fuel will come from natural gas. “This will increase global carbon emissions by 215-220 million tons through 2030, equivalent to 0.6% of global energy emissions.”
The analysis shows that electricity demand in AI data centers will increase by 165% between 2023 and 2030.
More from Electricity Wars
- Data Centers In The Middle Of Nowhere
- AI And AC Drive Up Electricity prices By 25%
- Air Conditioning Is Future Of Electricity
- Electricity Gets Ugly
