On the 14th of this month local time, research firm Gartner delivered a set of numbers to sober up the AI infrastructure revelers: between 2026 and 2030, total global investment in artificial intelligence infrastructure is expected to surge to $7 trillion (approximately RMB 46.86 trillion at the current exchange rate). Yet by 2030, these expensively built assets will generate at most $3.6 trillion in revenue—in other words, by then, less than half of the money invested will have been recouped.
Senior Director Analyst Will Sommer ran an even harsher calculation: factoring in financing costs and a 12% return on investment, the revenue actually needed to offset the $7 trillion investment would be as high as $8.5 trillion. And according to Gartner's revenue forecast, a gap of roughly $5 trillion lies in between. That is not the end of it—Gartner judges that sustained AI infrastructure investment may not reach the 12% return threshold until 2032; before then, more than $4 trillion in assets will face impairment risk.
The worst hit may be the "Neoclouds" caught between the giants—these emerging AI cloud service providers. Gartner predicts that by 2030, more than 80% of Neoclouds will cease independent operations. The crux: as chip iteration cycles accelerate and overall industry spending growth remains limited, these smaller cloud providers can neither spread costs across hardware upgrades nor compete with tech giants that own their own chips and customers, and in the end most will be forced out before they can break even.