The AI Spending Boom Could Come At A Huge Cost With Trillions at Stake, Report Warns

The artificial intelligence spending boom could leave trillions in infrastructure investment exposed to potential losses by the end of 2030, according to a new report.

The warning comes from Gartner's latest research, which estimates that AI providers will spend nearly $1 trillion on infrastructure this year and invest another $6 trillion between 2027 and 2030.

Much of that money is going toward data centers, advanced semiconductors and computing systems needed to train and operate increasingly sophisticated AI models.

But Gartner expects the revenue generated from enterprise and consumer AI services to fall well short of the amount required to deliver profitable returns on all that investment.

AI data center owners would need to generate a cumulative $8.5 trillion in revenue through 2030 to cover announced capital spending, financing costs and a 12% return on invested capital, the research firm estimates.

Revenue attributable to those data centers is projected to reach no more than $3.6 trillion over the same period, leaving a gap of approximately $4.9 trillion between the two figures. The 12% return is an assumption used in Gartner's model, rather than a guaranteed return or an industrywide requirement.

"In 2026, AI providers are expected to spend nearly $1 trillion on AI infrastructure," said Will Sommer, a senior director analyst at Gartner, adding that projected revenue would be insufficient to support profitable returns on all planned capital expenditures.

The forecasts do not mean the industry has already lost $3.5 trillion. The figure represents infrastructure spending that could become vulnerable to impairment, an accounting adjustment that reduces the recorded value of assets when their recoverable value falls below what a company has assigned to them.

Gartner estimates that continued spending at the current pace could leave more than $4 trillion in asset value exposed to potential write-downs through 2032. Under that scenario, AI infrastructure owners would not achieve the research firm's assumed 12% return on invested capital until the end of 2032 at the earliest.

The financial challenges could be particularly significant for neocloud providers, a newer group of companies specializing in renting high-performance computing capacity for AI workloads.

Unlike major technology companies with established cloud businesses and multiple revenue streams, these providers rely more heavily on demand for AI computing services to recover the cost of their infrastructure.

Gartner predicts that more than 80% of neocloud providers will cease independent operations by 2030 because of financial pressures and difficulties distinguishing their services from competitors.

That forecast concerns their ability to remain independent businesses. It does not necessarily mean that 80% will enter bankruptcy or stop providing services altogether.

The challenge differs for hyperscalers such as Amazon, Microsoft and Google, which operate large cloud platforms alongside other established businesses.

Gartner said their diversified revenue sources could help sustain AI infrastructure investment for longer periods, while specialized providers face greater pressure to generate returns directly from their computing services.

The warning follows several years of rapid growth in demand for computing resources used to develop and run artificial intelligence systems.

In an August forecast, Gartner projected that worldwide spending on AI-optimized infrastructure-as-a-service would grow 96.4% in 2026 to $42.3 billion, compared with $21.5 billion in 2025. Spending on these services is expected to reach $66.1 billion in 2027, an increase of 56.5%.

The firm attributed the expansion to continued demand for training large language models and the growing use of AI in business applications. It also identified a shift in the way companies use computing capacity.

Spending on AI inference, the process of running trained models to generate responses and perform tasks, is forecast to reach $23.3 billion this year, surpassing the $19 billion spent on model training.

Inference is expected to account for 55% of AI-optimized IaaS spending in 2026 and 59% in 2027. The growth reflects businesses moving beyond developing AI models toward using them in customer-facing products and everyday operations.