The extraordinary scale of America’s artificial intelligence investment boom has raised concerns that businesses could be sacrificing other projects to finance AI infrastructure. Goldman Sachs, however, believes the evidence of widespread displacement remains limited, even as AI spending approaches $600 billion and absorbs a growing share of corporate investment and debt issuance.
AI spending reaches significant share of U.S. investment
Goldman Sachs analyst Jessica Rindels expects U.S. AI investment to total almost $600 billion during 2026, “equivalent to nearly 2% of US GDP.”
The sector has accounted for more than 10% of business fixed investment in recent quarters, demonstrating how rapidly artificial intelligence has become a major component of corporate capital expenditure.
This concentration naturally raises concerns that spending on AI could come at the expense of factories, conventional technology upgrades or other corporate projects.
The heavy reliance on imported technology equipment also means the headline investment figures do not translate fully into domestic GDP growth.
Big Tech has cut buybacks rather than investment
For the largest technology companies, Goldman sees little evidence that AI infrastructure spending is forcing significant reductions in other capital projects.
Hyperscalers have partly financed their enormous AI programmes by allocating less cash to share repurchases.
They have also shown a willingness to tap debt markets.
Goldman said these companies have been “willing to borrow and appear undeterred by high interest rates.”
That financial flexibility allows the largest AI spenders to continue building infrastructure without necessarily making equivalent reductions elsewhere.
Corporate AI users show clearer evidence of displacement
Businesses consuming AI products and services present a different picture.
Goldman’s survey indicates that the absolute cost of adopting AI remains relatively small for most companies.
Yet around two-thirds of those expenses are being financed by reducing spending elsewhere.
This suggests AI adoption is already reshaping corporate budgets, although the amounts involved are not yet large enough to create a significant macroeconomic impact.
Data-centre construction reaches 9% of market
Data centres represent one of the clearest physical manifestations of the AI investment boom.
Goldman estimates that they now account for roughly 9% of private nonresidential construction expenditure.
Such rapid expansion might ordinarily be expected to create competition for labour, construction materials and other resources.
However, the data-centre boom has coincided with declining investment in subsidised manufacturing plants, helping offset some of the pressure.
As a result, Goldman has identified “only limited signs of crowd-out nationally.”
AI takes almost quarter of investment-grade issuance
The AI boom is also transforming corporate credit markets.
Financing associated with artificial intelligence now represents nearly one-quarter of investment-grade issuance, according to Goldman.
Despite this rapid increase, borrowing conditions for companies outside the AI ecosystem have not deteriorated significantly.
The bank said spillover effects “look limited so far,” pointing to non-AI credit spreads that remain near historically low levels.
That suggests debt investors have so far been capable of absorbing enormous AI financing requirements without materially restricting access to capital elsewhere.
Goldman estimates only $50 billion of additional crowding out
Goldman’s overall conclusion is that the economic consequences of the AI capital expenditure boom need to be viewed with more nuance.
The headline investment numbers are enormous, but imports reduce their direct contribution to domestic economic growth.
Meanwhile, evidence that AI is materially depriving other industries of financing, labour or construction resources remains relatively weak.
Goldman estimates approximately $50 billion of incremental crowding out during 2026, a relatively modest figure compared with the almost $600 billion expected to be invested in AI.
The bank therefore concluded that AI’s boost to GDP and the amount of other investment it displaces “are smaller than often thought.”
Leave a Reply