What Risks the US Economy Faces if AI Data Centre Boom Slows

AI is no longer just a tool for innovation – it's now a key component of the US economy. 

From major cloud platforms to chipmakers and hyperscalers, capital is flowing heavily into AI infrastructure.

At the centre of this investment surge are data centres. These facilities are now not only essential for powering generative AI, but they are also holding up the US economy – and that may be creating a problem.

Economists are beginning to question what happens if the AI investment cycle slows. 

As the traditional growth engines of the economy falter, spending on data centres and compute infrastructure is doing much of the heavy lifting. If that tailwind fades, the entire economic outlook could change.

Data centre spending takes centre stage in GDP growth

AI is driving vast capital expenditure, with hyperscale data centres forming the backbone of this surge. Companies including Microsoft, Amazon, Alphabet and Meta – the four largest spenders – are expected to invest US$344bn this year, according to Bank of America. That equals 1.1% of US gross domestic product (GDP) and marks a rise from US$228bn last year.

Barclays estimates that this capital flow into software, compute equipment and data centres adds one percentage point to GDP growth in the first half of 2025. 

With total GDP growth at 1.6% over that period, this suggests that AI infrastructure alone contributes 0.8% – or half the total.

The level of dependence is causing concern. Peter Berezin, Chief Global Strategist at BCA Research, says: “It’s certainly plausible that the economy would already be in a recession” without the AI boom.

Strip away this investment and the economic picture looks weaker.

Deutsche Bank notes that private business investment outside of AI-related categories has been flat since 2019. Traditional commercial construction – including offices and retail spaces – is in decline. Job creation has slowed and unemployment is climbing.

Stephen Juneau, economist at Bank of America, says: “It’s the only source of investment right now.”

Data centres reshape construction and supply chains

The scale of data centre development is reshaping the US construction landscape. Turner Construction now reports that data centres make up around 35% of its project backlog – a sharp increase from 13% five years ago. 

These sites require large workforces to build. Ben Kaplan, Managing Director of the advanced technology group at Turner Construction, says each data centre project involves anywhere from 100 to 5,000 people.

The surge is straining supply chains. Lead times for essential components like switchgear and electrical generators are stretching by several months. “Every element of the supply chain is being stressed right now,” says Kaplan.

While much of the hardware spend goes on imported chips – especially those made by Nvidia – domestic infrastructure still sees major investment. Nvidia expects US$65bn in sales in the fourth quarter of this year, beating forecasts

Meanwhile, Bank of America projects that the four big tech firms will raise capital expenditures again in 2026 to US$404bn, though growth in spend is likely to slow.

The demand from AI and data centre operators has also pushed companies such as Oracle and CoreWeave into aggressive borrowing. 

Oracle now holds over US$100bn in debt, after issuing US$18bn in bonds to fund AI infrastructure. CoreWeave, which rents data centre space and compute power to technology firms, is also taking on heavy debt to meet demand.

If revenue fails to keep up with borrowing, this could feed instability into credit markets.

Market risks rise as valuations climb

High stock valuations add another layer of risk. If current earnings projections do not hold, tech stocks could correct – and that would have wider economic effects. 

The S&P 500 index fell around 2% last week, as fears of an AI bubble started to dent investor confidence.

This matters for consumption. JPMorgan Chase calculates that rising prices in AI equities have added US$180bn in consumer spending over the past year.

That figure reflects the wealth effect – when households spend based on paper gains in share prices. A reversal would likely suppress consumer activity.

Jonathan Millar, Senior US Economist at Barclays, warns of the impact if markets fall: a 20% to 30% correction in stock prices could cut GDP growth by 1% to 1.5% over the course of a year. 

A pause in AI investment could trim another 0.5% If it collapses, a full point could be lost.

Peter Berezin summarises the threat: “If you take a fragile labour market and you kick it with a capex bust, you’re probably going to get a recession out of it.”

As the US economy leans more heavily on data centre development and AI infrastructure, the risks of over-dependence grow. 

Without continued momentum, the knock-on effects could reach across labour markets, consumer spending, credit and overall growth.