Hyperscalers have poured so much money into building as much AI infrastructure as possible—and as quickly as possible—that investment from a handful of companies is expected to reach $1 trillion a year soon.

Meanwhile, the housing market has been largely frozen since the COVID-era boom ended in 2022, when the Federal Reserve embarked on an aggressive rate-hiking campaign to rein in inflation.

“We’re seeing a pivotal shift in the US economy: investment is shifting away from residential investment and towards computers,” Adam Shapiro, vice president at the San Francisco Fed, posted on LinkedIn recently.

He pointed out that inflation-adjusted spending on information processing equipment, which includes data centers and computer hardware, now exceeds residential investment.

According to data from the Bureau of Economic Analysis, real private residential fixed investment was $748 billion in the second quarter, down 18% from an early 2021 peak. During that same span, spending on information processing equipment has soared 51% to $752 billion.

“The AI investment boom is massive,” Shapiro added.

He also noted that residential investment is more sensitive to borrowing costs, which have gone up alongside Treasury yields. The benchmark 30-year mortgage rate is nearly 7% as the 10-year bond yield has hit the highest level since 2007.

By contrast, AI investment has been less sensitive to interest rates, even as hyperscalers have started issuing more debt to supplement drawdowns of their cash piles. Google parent Alphabet, for example, even reported negative cash flow earlier this year.

Treasury Secretary Scott Bessent has also highlighted the eagerness with which AI companies are offering debt—no matter the cost of borrowing.

“We are also seeing big corporate issuance. And a lot of that corporate issuance, I would say, is almost yield-agnostic, because the build-out for AI, the returns on that, the companies believe they’re going to be so high. They don’t really care what they’re paying,” he said recently.

The onslaught of AI spending is expected to keep ramping up. S&P Global estimated last month that capital expenditures from Alphabet, Amazon, Microsoft, Meta, Oracle, and SpaceX will exceed $1.3 trillion in 2027, up from a projected $870 billion in 2026 and $470 billion in 2025.

The ratings firm added that the industry’s capex is growing faster than revenue, warning that the aggressive build-out could lead to overcapacity if future demand doesn’t pan out as expected.

S&P sees 2028 as an inflection point, with revenue accelerating and capex flattening. Until then, however, operating cash flow from the six hyperscalers will collectively be negative in 2026 and 2027, it added.

The breakneck speed of AI development has also generated immense political backlash that’s rippling through the midterm election season.

A new NBC News poll found that 64% of registered voters said they’d be less likely to support a candidate who’s in favor of building a data center in their community.

AI has also fed into the cost-of-living crisis as consumers grapple with higher electricity bills and prices for new smartphones and PCs. That’s added to angst with home prices remaining out of reach for many Americans.

While increased housing supply would ease pressure on prices, availability has been limited. The supply of existing homes has been constrained by the “lock-in” effect of homeowners with low mortgage rates reluctant to give them up amid today’s high rates.

And new construction has been weak as high rates weigh on demand while building costs grow. Housing starts fell 2.6% in August to an annualized pace of 1.275 million, led by a drop in multifamily projects. While single-family homes rose, permits fell, signaling muted activity in the future.

Meanwhile, the National Association of Home Builders reported builder sentiment fell to its lowest level in a year.

“The big picture remains that elevated and rising borrowing costs are holding developers back, supporting our view that the downward trend in housing starts has further to run,” Capital Economics said in a recent note.