Gartner Says $3.5 Trillion of AI Spending Could Be Written Down, and Most Neoclouds Won't Survive on Their Own

Research firm Gartner has put hard numbers on the AI bubble debate, and they are sobering for anyone who owns data centers. More than $3.5 trillion of AI infrastructure spending could face write-down risk by 2030, Gartner warns, and more than 80% of neocloud providers, the specialized AI cloud companies like CoreWeave and Nebius, will stop operating independently by the end of the decade. The forecast landed on the same day those stocks sold off hard on worries about rising borrowing costs.

The math doesn't add up

The core of Gartner's warning is a revenue gap. AI providers are projected to spend nearly $1 trillion in 2026 and another $6 trillion through 2030. To cover that spending, pay the financing costs and earn a 12% return, Gartner estimates data center owners would need $8.5 trillion in cumulative revenue. But it sees no more than $3.6 trillion of AI revenue that can be attributed to data centers.

Even the revenue that exists won't all reach the owners of the hardware. Gartner expects infrastructure owners to earn less than half of AI revenue, with much of the market running on asset-light business models, such as software and AI services that rent computing rather than own it. As a result, Gartner does not expect the targeted 12% return to be reached before late 2032.

Why neoclouds are most exposed

Neoclouds face the toughest math. Gartner projects spending on AI cloud infrastructure will grow from $66 billion in 2027 to $174 billion in 2030. That is strong growth, but possibly not fast enough for these companies to earn back their investment before today's generation of AI chips becomes obsolete. GPUs lose value quickly as newer models arrive, so a data center that takes too long to pay off may never fully recover its cost.

That is why Gartner expects more than 80% of neoclouds to lose their independence by 2030, whether through acquisitions, mergers or failures, as cost pressures build. The big diversified hyperscalers, such as Microsoft, Alphabet and Amazon, are in a stronger position because they have other profitable businesses and far more financial flexibility to ride out a downturn.

A warning on a rough day

The report adds weight to a week of warnings about how the AI build-out is being financed. On Tuesday, the CEOs of Goldman Sachs, JPMorgan and Wells Fargo flagged the risk of a painful reset in debt-fueled AI spending. On Wednesday, CoreWeave fell nearly 16% after a chart of its rising borrowing costs went viral, dragging Nebius, Oracle and AI-hosting miners down with it. Gartner's forecast suggests investors' concerns are about more than one bad day.

What to watch

• Hyperscaler capex guidance. Microsoft, Alphabet, Amazon and Meta report later this month. Any slowdown in their spending plans would ripple through the whole supply chain.
• AI revenue growth. Faster-than-expected growth at AI software and service companies would help close Gartner's revenue gap.
• Neocloud financing. Borrowing costs and new funding deals for CoreWeave, Nebius and peers will show how much runway they have.
• Consolidation. Mergers or acquisitions among neoclouds would be an early sign Gartner's prediction is playing out.