Gartner: 80% of neoclouds will vanish as AI math fails to add up

Gartner's math says the AI build-out is running on a deficit. The research firm projects that 80 percent of neocloud operators, the GPU-as-a-service providers that rent out AI compute by the hour, will not survive as independent companies beyond 2030. The warning comes from a report titled 'Prepare for AI Market Consolidation: Capital Write-Downs Could Exceed $4 Trillion by 2032,' which concludes that AI will not generate enough sales to profitably cover planned AI capital spending before 2032 at the earliest.

The gap is staggering. Gartner senior director analyst Will Sommer said AI providers are expected to spend nearly $1 trillion on AI infrastructure in 2026, then invest an additional $6 trillion between 2027 and 2030. To make that capital work at a 12 percent return on invested capital, the bar many companies use for acceptability, operators would need to generate $8.5 trillion in revenue over the period. Gartner forecasts the actual enterprise and consumer AI revenue attributable to AI datacenters will reach no more than $3.6 trillion. That leaves a roughly $4.9 trillion hole between the revenue needed and the revenue expected, and it is the reason Gartner expects a wave of write-downs as asset values can no longer be recovered.

Write-downs are not just an accounting exercise. Newer GPUs displace older hardware quickly, and the value of datacenter assets can deteriorate faster than depreciation schedules assume. When returns fall short, operators are forced to recognize that the infrastructure is worth less than what they paid for it. Gartner believes neoclouds will be the hardest hit because their entire business model is built on renting AI compute, with no diversified revenue streams to cushion the blow. Hyperscalers, by contrast, can absorb the pain because their businesses extend well beyond AI services.

The warning lands in a market already getting nervous. Last month, SoftBank Group faced higher borrowing costs as it sought additional funding for OpenAI, a sign that lenders are starting to price in AI risk. Jefferies issued a report this week saying developers such as OpenAI and Anthropic face profitability threats from open source and Chinese models that undercut them on price. Management consultant Bain & Company estimated last month that the AI industry needs to generate $6 trillion in annual revenue by 2031 to keep funding the infrastructure required to meet anticipated demand, and it forecast a shortfall unless the industry develops products and use cases not yet identified.

Macro voices are chiming in too. Both the International Monetary Fund and the Bank for International Settlements have voiced concerns about an AI bubble bursting and the effects rippling across the global economy. The pattern is familiar: a technology with genuine promise attracts enormous capital, supply outruns demand, and the weakest players get consolidated or wiped out. McKinsey warned last year that neocloud companies need to move up the AI stack and diversify to avoid being commoditized, and Gartner's forecast suggests that warning was not heeded broadly enough.

For executives and boards, the strategic stakes are clear. The AI infrastructure build-out is being treated as a land grab, but the revenue side of the equation is not keeping pace. Companies that are pouring capital into datacenters, GPUs, or AI services need to stress-test their plans against Gartner's numbers: if the market can only support $3.6 trillion in revenue against $7 trillion in spending, some of that capital will not earn its cost of capital. The question is not whether there will be a shake-out, but who will be on the buying side and who will be on the selling side when it comes.