Global capacity could triple to 219GW if AI proves useful, firm says, but only if the technology finds real-world utility.
Data centers could need trillions of dollars of investment over the next five years to keep up with demand, management consulting firm McKinsey has claimed, but only if AI finds real world value.
“Our research shows that by 2030, data centers are projected to require $6.7 trillion worldwide to keep pace with the demand for compute power,” the company said, a number it called “staggering.”
In a report published this week titled The cost of compute: A $7 trillion race to scale data centers, the company said global demand for data center capacity could almost triple by 2030 – to 219GW – with about 70 percent of that new demand coming from AI workloads.
Within that $6.7t figure, data centers equipped to handle AI processing loads are projected to require $5.2t in capital expenditures, while those powering traditional IT applications are projected to require $1.5t in capex.
If that $5.2t for AI investment comes to fruition, the firm estimates approximately 15 percent ($800 billion) will flow to builders for land, materials, and site development. Another 25 percent ($1.3t) will be allocated to energy suppliers for power generation and transmission, cooling, and electrical equipment. The largest share of investment, 60 percent ($3.1t), will go to technology developers and designers, which produce chips and computing hardware for data centers.
The firm noted, however, that “a lack of clarity” about future demand makes precise investment calculations difficult: “Our analysis is built on thoroughly researched hypotheses, but there are critical uncertainties that cannot yet be quantified," the report said.
Two key uncertainties it noted were whether or not we’ll see AI use cases turn into real business impact that generate value, and if increases in training and inference efficiency might lead to reduced infrastructure capacity demand.
McKinsey said the challenge for investors in the space is formidable: “Deciding how much capital to allocate to which projects, all while remaining uncertain of how AI’s future growth and development will impact compute power demand. Overinvesting in data center infrastructure risks stranding assets, while underinvesting means falling behind.”
The true figure could be much higher or lower, depending on various factors, but is still predicted to be in the trillions. The company looked at three scenarios – accelerated demand, continued momentum, and constrained momentum. Factors that will influence those figures include mass adoption of generative AI, enterprise integration of AI, competition between companies, and geopolitics.
At the constrained lower end of the company’s forecast, 78GW of new capacity could come online by 2030, requiring $3.7t in new investment. At the accelerated higher end, the company predicts 205GW of capacity could be added, totaling $7.9t of investment.
