The AI investment boom has unleashed a wave of demand for capital just as savings are declining and interest rates remain high. As competition for financing intensifies, rising borrowing costs could squeeze non-AI firms looking to finance new factories, equipment, and projects.
LONDON—Wall Street analysts keep revising their forecasts for AI-related capital expenditures upward, with some estimating that major tech firms—including Amazon, Microsoft, Alphabet, Nvidia, and Meta—could spend as much as $1.4 trillion by 2027. Gartner, a research and advisory firm, projects that global AI spending will reach $2.5 trillion this year.
The staggering amount of capital now being earmarked for data centers sprouting across the United States, advanced networking equipment, and electricity infrastructure raises a critical question: Is the AI buildout crowding out investment in other sectors? As Goldman Sachs economists Jessica Rindels and David Mericle—hardly anyone’s idea of modern-day Luddites—observed, there is a real risk that AI could starve other productive sectors of capital and become a drag on growth.
Indeed, the consensus among investors, as Apollo Global Management’s Chief Economist Torsten Slok recently noted, is that hyperscaler capital expenditure will reach roughly 3% of GDP annually between 2027 and 2029. That is sharply up from 0.3% of GDP in 2019 and 1.4% in 2025.
AI industry leaders and investors argue that the current AI boom could deliver enormous productivity gains, transforming the global economy and leaving everyone better off. Like previous technological revolutions, AI promises to restructure entire industries, from defense and manufacturing to education. It could also reduce costs in sectors, especially education and health care, where efficiencies have long been elusive.
AI-related investments have already become a major driver of US economic growth. Morgan Stanley estimates that capital expenditures on the technology could add as much as 2.5% to US GDP growth this year and more than 3% in 2027.
But is the AI spending spree coming at the expense of other industries? In 2025, AI accounted for 65.4% of venture-capital deal value in the United States. With its voracious appetite for capital, there is a real chance that the industry could starve other promising fields, such as biotech and nanotechnology, of much-needed investment. Already in 2024, AI engineer Abhay Gupta warned that startups in suddenly less-glamorous areas like greentech and social impact investing were finding it much harder to attract VC capital because of the AI boom.
Crowding out is not confined to venture capital. It can also occur when fierce competition for capital pushes borrowing costs high enough to choke off financing for otherwise attractive projects or sectors.
There are already signs of this dynamic in capital markets, the most obvious one being that borrowing costs, as measured by real interest rates, have been rising. The recent spike in US Treasury yields could partly reflect the growing demand for financing fueled by the AI investment boom; new Fed Chair Kevin Warsh’s hawkish approach to inflation no doubt also contributed. For the first time in more than 30 years, capital is being repriced as demand begins to exceed supply. If AI is driving that shift, the result could be a massive crowding-out effect.
To put the sheer scale of capital flowing into AI in perspective, relative to US GDP, the AI super cycle ranks ahead of previous investment booms, including the mid-19th-century railroad boom, the construction of the US interstate highway system from 1955 to 1970, and the Apollo space program.
The range of financing tools tapped to satisfy this demand is equally striking. Virtually every financial instrument and innovation developed over the past few decades is being used to fund the AI buildout, from plain-vanilla investment-grade bonds and securitizations to private credit, leasing-style structures, and retail funds.
Naturally, such a vast amount of debt (and equity) issuance has fueled concerns about circular financing deals in which major AI players, including chipmakers and cloud providers, fund startups in the AI space. Those startups then spend that capital on hardware and computing power from the very same providers.
This circular financing structure poses risks to the financial system and the real economy, particularly at a time of rising bond yields, when leverage among AI businesses could spark a financial crisis. Bank of England Governor Andrew Bailey recently warned of a potential “blow-up,” or collapse of the AI investment bubble, that could trigger a major global financial downturn.
The Big Squeeze
Compounding this surge in demand for capital is a decline in savings. Since OpenAI released ChatGPT in November 2022, the US household savings rate has fallen from 3.8% to 3% while AI investment has surged. This stands in stark contrast to the situation during then-Federal Reserve Chair Ben Bernanke’s tenure, from 2006 to 2014, when a savings glut coincided with a private sector dominated by businesses with relatively modest capital needs, keeping interest rates low. Today, a capital-intensive industry is dominating investment flows just at a time of declining savings and rising interest rates.
This increases the risk of crowding out investment in other industries. The private sector has a finite pool of capital, and its size is largely determined by how much people are willing to save. The more of those savings AI absorbs, the less there is for everything else, including factories and other investments that require large amounts of capital.
Debt markets provide an early indication of this strain. In the first eight months of 2026, the five largest hyperscalers issued $132 billion in debt, compared with an annual average of $35 billion in debt between 2020 and 2024. And that doesn’t even include financing for data centers, chipmakers, and utilities. Vanguard estimates that AI-related debt issuance could total $300–570 billion for the full year.
Rising capital costs signal that demand is outpacing supply. With so much capital flowing to data centers and AI infrastructure, higher interest rates make it more expensive for other sectors to raise money in the bond market. As AI-driven demand for capital rises, borrowing from other industries must fall in order for the market to return to balance. When the supply of capital is inelastic, some demand inevitably goes unsatisfied. If this actually happens, the economy may well experience the “innovation winter” that the Eurasia Group warned about in 2019 as a potential consequence of capital starvation.
Another indication of crowding out is the widening gap between the performance of AI-related stocks and the rest of the market. From 2023 through 2025, the “Magnificent Seven”—Alphabet, Amazon, Apple, Meta, Microsoft, Nvidia, and Tesla—outperformed the S&P 500 by a wide margin, at times accounting for as much as two-thirds of the index’s total gains.
Together, these seven companies now make up more than one-third of the S&P 500’s total market capitalization. RBC Wealth Management has called this concentration the “Great Narrowing,” describing it as even more pronounced and entrenched than during the dot-com era, which of course ended in a crash. Although RBC acknowledges that concentration does not necessarily mean a bubble, the risks are real, ranging from the damage that a poor earnings report from just one company could inflict on the overall index to the more serious threat to the economy if sentiment toward AI suddenly turns negative.
Capital is also not the only resource AI is competing for. Its insatiable demand for water, electricity, land, and materials could drive up prices and leave other sectors to compete for increasingly scarce supplies.
Recent company earnings calls suggest that this is already happening in some areas. Intel, for example, has indicated that semiconductor supply is being directed toward AI and away from PCs. Elevator manufacturer Otis and homebuilder Lennar have pointed to skilled-labor and construction-capacity shortages that are delaying projects. IBM has noted that IT budgets are shifting toward AI and away from conventional hardware and software, while Micron and SK Hynix have reported memory shortages constraining PC and smartphone production.
Crowding Out or Crowding In?
Despite considerable financial and anecdotal evidence, the claim that AI is crowding out other investment is far from settled. There are at least three reasons to question this narrative, one of which suggests that the AI boom could be having precisely the opposite effect.
For starters, there is no clear evidence of crowding out in the macroeconomic data. AI may be consuming a greater share of investment, but that has not been offset by a decline in investment elsewhere. In fact, data from the Bureau of Economic Analysis show that gross private investment as a share of US GDP has remained broadly stable.
Second, higher interest rates could reflect a repricing of financial assets rather than crowding out. The 30-year Treasury yield has risen by more than 33 basis points since the start of the year, but that increase has been attributed largely to the Iran war and renewed inflation concerns following the spike in energy prices. With so many forces driving up rates, it is hard to say with any certainty how much of this trend reflects the impact of AI.
The third counterargument draws on the so-called Jevons paradox: when a new technology makes a resource more productive and efficient, the resulting cost declines can strengthen, rather than weaken, demand for the resource. By this logic, AI could increase productivity, lower costs, and boost demand for a wide range of goods and services, prompting companies to invest more in response. AI investment could thus become a catalyst for investment in other sectors—power generation and utilities, financial services, and cloud computing—rather than sucking capital away from them.
For now, AI appears to be crowding out investment at the margins and not across the board. It is clearly diverting capital and resources away from other sectors, but there is not yet enough evidence to conclude that it is dragging down overall investment or economic growth.
About the Author:
Dambisa Moyo is an international economist, is the author of Edge of Chaos: Why Democracy Is Failing to Deliver Economic Growth – and How to Fix It (Basic Books, 2018).