The global race to build the backbone of artificial intelligence is creating an unexpected ripple effect that is being felt far from the server farms of Northern Virginia or the tech hubs of the West Coast. It is showing up in the cost of financing a new apartment complex in the suburbs and the price of steel for a residential tower in Manhattan.
Scott Rechler, the CEO of RXR and a member of the board of the Federal Reserve Bank of New York, is sounding an alarm on what he describes as a massive displacement of capital. Speaking at a recent Trepp conference, Rechler argued that the sheer scale of corporate borrowing to fund AI infrastructure is driving up interest rates and starving the multifamily housing sector of the affordable capital it needs to address the national housing shortage.
For those of us who have watched the markets transition from the low-rate environment of the 2010s to the current volatility, the mechanics are familiar but the scale is unprecedented. When corporate debt issuance for digital infrastructure surges—potentially rivaling or exceeding U.S. Treasury issuances in certain forecasts—it creates a “crowding out” effect. As the demand for loanable funds spikes to build GPU-heavy data centers, the cost of borrowing rises for everyone, including the developers trying to build the homes Americans actually need.
The Speculation Risk in Digital Infrastructure
The core of the problem, according to Rechler, is how the market is categorizing these AI projects. Investors are treating data center developments as “infrastructure”—a category typically reserved for low-risk, stable assets like toll roads or water treatment plants. In reality, Rechler argues, these projects carry significant execution risk.
While traditional infrastructure is often built with long-term government contracts or guaranteed usage, a growing number of data centers are being built “on spec.” So developers are breaking ground without secured tenants, gambling that the demand for AI compute will remain insatiable enough to fill the space upon completion. If the AI bubble bursts or the technology evolves to require different hardware footprints, the industry could be left with “stranded assets”—massive, energy-hungry shells that no one wants to rent.
This speculative fever doesn’t just risk a future crash; it creates immediate pain for other sectors. Because AI companies are often backed by massive cash reserves or aggressive venture capital, they are outbidding traditional developers for essential resources. This isn’t just about money—it’s about the physical components of construction.
The War for Labor and Equipment
The competition for specialized labor and industrial equipment has reached a fever pitch. From electrical engineers to high-capacity transformers, the supply chain is being cannibalized by the AI boom. Rechler noted that the inflationary pressure on construction costs for commercial and residential projects is being driven upward by tech giants who can afford to pay any price to stay ahead in the AI arms race.
RXR is feeling this pressure firsthand. The firm is currently planning a supertall tower at 175 Park Avenue in New York City, a project that exemplifies the modern struggle for materials. To mitigate the risk of supply chain collapses, Rechler revealed that RXR has already purchased turbines and generators for the tower before even breaking ground—a defensive move that would have been unthinkable in a more stable market.
| Factor | Traditional Infrastructure | AI Data Center Spec-Builds |
|---|---|---|
| Tenant Security | Long-term contracts/Government backed | Often speculative (no secured tenant) |
| Risk Profile | Low volatility, steady cash flow | High execution and technology risk |
| Capital Impact | Predictable borrowing patterns | Aggressive debt issuance, raises rates |
| Supply Chain | Standardized materials | High demand for specialized power/cooling |
Parallels to the Dot-Com Bubble
Rechler is no stranger to market cycles. He famously sold his real estate business in 2007, just before the global financial crisis wiped out much of the sector and only returned to buying Manhattan office buildings in 2022. Now, he sees parallels between the current AI investment wave and the dot-com bubble of the late 1990s.
He pointed to an estimate from IBM CEO Arvind Krishna, suggesting that only about 35 percent of current AI companies may survive the next two to three years. If a significant portion of these companies fail, the debt they’ve issued to build their infrastructure could become a systemic weight on the financial markets, much like the commercial real estate crisis of 2008 or the tech crash of 2000.
Yet, the paradox for any modern business leader is that you cannot afford to ignore the technology, even if you fear the bubble. RXR is actively integrating AI into its own operations to maintain a competitive edge. The firm currently uses AI to monitor its 30 million-square-foot portfolio, employing algorithms to scan news reports and government filings for any data that might impact property values. They have also deployed drone imagery analysis to track construction progress in real-time, reducing the need for manual site visits and improving oversight.
The Broader Economic Impact
The implications of this shift extend beyond the balance sheets of real estate firms. When corporate borrowing to fund AI competes directly with government bond issuance, it forces the U.S. Treasury to offer higher rates to attract investors. This creates a feedback loop: higher Treasury yields lead to higher mortgage rates, which in turn makes it more expensive for multifamily developers to finance new housing units.
For the average renter or homebuyer, the “AI revolution” manifests not as a smarter chatbot, but as a higher monthly rent check because the developer of their building had to pay 2% more on their construction loan due to the debt appetite of a data center developer three states away.
Disclaimer: This article is for informational purposes only and does not constitute financial, investment, or legal advice.
The industry now looks toward the next Federal Open Market Committee (FOMC) meeting and upcoming Treasury issuance schedules to see if the cost of capital begins to stabilize or if the “crowding out” effect continues to intensify. As AI companies move from the procurement phase to the operational phase, the market will discover whether these speculative data centers can generate the cash flow necessary to service the massive debts they’ve incurred.
Do you think the AI infrastructure boom is a sustainable evolution or a bubble waiting to burst? Share your thoughts in the comments or share this story with your network.
