Wall Street Pitches Data Centers as Prime Real Estate, But Risks Mount

Data centers have become a focal point of division in the United States. While Wall Street investors are eagerly backing artificial intelligence as the future engine of economic growth, everyday citizens are increasingly opposing the construction of these massive facilities. As infrastructure powering major AI models such as Google’s Gemini, Meta’s Muse, Anthropic’s Claude, and OpenAI’s ChatGPT expands across the country, financial firms are attempting to market AI infrastructure as a diversified real estate asset class.

Although these alternative investment opportunities have primarily targeted institutional clients like pension funds, they have begun to reach retail investors. Blackstone has led this charge with the launch of the Blackstone Digital Infrastructure Trust (BXDC), a newly formed real estate investment trust listed on the NYSE earlier this year.

CEO Nick Pell stated in a May interview on CNBC that the firm viewed the data center market as an opportunity to establish a public market presence for stabilized assets. He emphasized that focusing on mature markets like Northern Virginia and Dallas represents the “lowest risk” strategy, citing a $300 billion total addressable market. The trust priced its initial public offering at $20 per share for 87.5 million shares. However, the fund has since declined by approximately 16%, trading under $17 as of Thursday. Blackstone declined to provide further comment on the performance.

Other major players in the sector include Equinix and Digital Realty, both of which operate as REITs. While they have seen stronger long-term performance, their shares have stagnated since the BXDC launch. The broader REIT sector has outperformed typical expectations during periods of rising interest rates this year, although many gains occurred before bond market stress intensified in August.

Pell described the opportunity as “massive,” projecting that the total addressable market for the business could exceed $1 trillion in the coming years. AI-driven construction has notably propped up national construction spending even as other sectors face declines.

Competition is heating up. Blue Owl is reportedly considering a public REIT valued at up to $6.5 billion, which would consolidate its existing portfolio of over 130 data centers across 32 global markets, representing more than $18 billion in assets. Co-CEO Marc Lipschultz has touted data centers as one of the strongest long-term investment opportunities in decades, highlighting contracts structured to protect investors even if tenants exit early. Meanwhile, Brookfield Asset Management launched Csquare as a standalone NYSE trading vehicle in July; its shares have dropped nearly 16% since debut. Unlike its competitors, Blackstone’s BXDC has not yet deployed capital into specific investments.

Despite the bullish financial pitch, the sector faces growing political headwinds. Gallup polling indicates that 70% of Americans oppose data centers in their area, a sentiment that crosses party lines. New York became the first state to impose a moratorium on new hyperscale data center approvals in July, followed by Texas in August, where Governor Greg Abbott halted new permits despite previously calling the state the AI “epicenter.” These political shifts pose risks not only to new projects but also to those already included in fund assumptions.

Financial repercussions are already visible. Oracle’s stock fell 4% in late September after it issued a force majeure notice regarding Project Jupiter, a New Mexico data center campus linked to the Stargate AI buildout. Oracle cited regulatory hurdles and local opposition, seeking to delay payments if the facility is not operational by 2028, though the company maintains the project remains on track. Globally, an Nvidia and Blackstone-backed Australian data center company recently canceled its planned IPO due to weak investor interest.

Broader market volatility associated with the AI trade has also impacted sentiment. Many AI-themed stocks declined after reports emerged that OpenAI’s revenue forecast was approximately $20 billion lower than previous projections.

Experts acknowledge the complexity of the sector. Sabur Mollah, a finance professor at Gettysburg College, noted that while high-quality facilities can provide stable rental income through long-term leases with major tech companies, the industry is capital-intensive and heavily reliant on electricity, cooling, and connectivity. He warned that excessive enthusiasm for AI could lead to overvaluation, alongside risks related to tenant concentration, technological obsolescence, and refinancing costs.

Risk advisers at IMA, an insurance brokerage, highlighted additional non-market dangers. Patrick Datz pointed out that while traditional locations like Virginia’s “Data Center Alley” were chosen for their low disaster risk, the push to build everywhere exposes assets to greater natural hazards. Rachel Nixon added that any failure to maintain promised uptime could trigger contractual problems. Despite these challenges, both Datz and Nixon maintain that data centers remain a fundamentally sound investment.