US Real Economy's Reliance on AI Data Center Construction Investments Resembles the 2001 Dot-Com Bubble
The New York Times has assessed that the US real economy's growing reliance on AI data center construction could be similar to the 2001 dot-com bubble burst, which triggered a stock market crash and a wave of fiber optic cable failures.
The NYT reported on the 27th, "The intensity of the AI investment wave evokes uncomfortable parallels to the last time, in 2001, when the tech industry poured billions of dollars into infrastructure to support new technologies with high expectations of future profits."
The report continued, "Following the stock market crash of 2001, triggered by the collapse of speculative dot-com companies, the telecommunications sector also collapsed, and the companies that borrowed to build fiber optic networks went bankrupt, triggering a global economic meltdown." Already, the New York Times reports, “OpenAI CEO Sam Altman raised eyebrows this month by saying the sector is ‘overexcited’ and ‘some players will lose a lot of money.’” The investment bank UBS, while generally positive on the industry, wrote in a note to clients that there may be some ‘indigestion’ about ongoing capital spending.
Investors are now almost certain that the downturn won’t be catastrophic.
The differences from the 2001 crash, the New York Times notes, include that “data centers are financed by a diverse group of lenders, reducing their exposure to one part of the banking system,” and that “current leases are typically long-term and difficult to escape, which can insulate landlords even if their wealthy tenants have to leave.” Investment bank UBS estimates that companies will spend $375 billion globally on AI infrastructure this year, 2025, and that figure will increase to $500 billion next year.
The New York Times reported, "Commerce Department data shows that investment in software and computer equipment, excluding AI data center buildings, accounted for a quarter of overall economic growth in the second quarter." The Times added, "AI infrastructure has overtaken office construction in the United States, and spending on data centers has surged in recent years, not even including the equipment that goes into them."
AI data centers are also attractive to traditional construction companies, which see an opportunity to shift from standard real estate development projects to a new, well-capitalized asset class.
Skanska, a major contracting firm, forecasts that data center construction will grow at a compound annual growth rate of 13.2% through 2029, a much faster rate than any other sector.
The American Cement Association, an industry group, estimates that the sector will require one million tons of cement in the coming years. Amrize, one of North America's largest building materials suppliers, has developed an "AI-optimized" concrete mix with a lower carbon footprint.
Amrize told the New York Times that the data center construction boom was a "bright spot" in the otherwise sluggish second quarter.
The AI data center construction boom has created a boom for electricians, engineers, and heavy equipment operators. While operational data centers typically employ only a handful of workers, the construction phase of an AI data center can employ thousands, creating a temporary tech labor boom.
This temporary boom is one reason why US construction employment remained stable in the second quarter, despite the stagnation of housing, office, and warehouse projects, which led to a recession.
Duos Technologies, which provides analytics and imaging for railroads and large-scale infrastructure, recently expanded into smaller data centers.
"Our business is well-positioned, both commercially and financially, to capitalize on the very hot demand resulting from the data center computing gold rush," Duos CEO Charles Ferry told the New York Times. Dennis Calvert, CEO of BioLargo, an environmental services company that sells large-scale battery systems, told the New York Times, "These data center managers and large AI providers have an insatiable need for energy and energy storage."
James Walker, CEO of NANO Nuclear Energy, which makes small nuclear reactors, said, "The opportunity for advanced nuclear power is stronger than ever, as the growth of data centers and climate mandates accelerate the demand for clean, reliable baseload power."
According to JLL, a real estate services firm, its North America Data Center Report, "$1 Trillion in Data Center Construction from 2025 to 2030," states that "over 100 gigawatts (GW) of capacity (colo + hyper) could be started or completed between 2025 and 2030." Investors are very willing to finance development, with 73% of projects under construction having been pre-approved. “If you think about data generation and data storage in general, it’s been growing at a rapid pace for decades and will continue to grow,” Andrew Batson, JLL’s head of data center research in the Americas, told the Times. “At some point, demand will naturally slow down, but that’s not in the short term. We expect the sector to grow about 20% annually until at least 2030.”
The biggest constraint on data center growth in the coming years will likely be the rising costs of water and electricity, but more likely the rising costs of water and electricity energy workers and technical equipment.
“The energy, water, workers and technical equipment needed to build and operate data centers are all getting more expensive,” the Times said. “At the same time, communities once eager to host data centers are sometimes unhappy about them, most recently the St. Charles, Missouri, city council imposing a year-long moratorium on new facilities due to concerns about drinking water contamination.” Eric Gaus, chief economist at Dodge Construction Network, which closely tracks new developments, said, "There's a lot of money being spent, but at some point, the cost of the fruit will bite. A growing number of local residents are saying, 'If we're going to put something here, we need to do more than just build it and leave,'" adding to the external factors driving the rising cost of data centers.
The New York Times reported, "Investors have been fueling the stock market by turning chip giant Nvidia's earnings report, released on the 27th, into a 'windfall' of optimism about the potential benefits of AI." They added, "While recent months have shown that AI spending is also boosting the real economy, at least this investment isn't yet driven by how companies are using AI technology."
Instead, the current stock frenzy stems from a surge in business activity, temporarily brightening the overall domestic economic figures, driven by the massive investment in data centers, semiconductor factories, and power supplies needed to build the computing power AI demands.