Wall Street is pitching data centers as a major real estate bet. The risks are piling up

Blackstone’s newly launched data-center REIT sold 87.5 million shares at $20 each in mid-May but is down roughly 16%, and it has not yet deployed capital. Investor risks include rising political opposition—70% of Americans oppose a data center in their area, according to Gallup—alongside power constraints, construction delays, high capital needs and limited liquidity in some private funds. Long-term demand for AI and cloud infrastructure supports the investment case, but New York and Texas have halted approvals for new hyperscale projects, and Oracle cited regulatory hurdles and local opposition in seeking a possible payment delay on a New Mexico campus while saying it remains on track.

Analysis

Investment view: The market is distinguishing operating capacity from a vehicle offering exposure to future capacity. BXDC’s lack of deployed capital makes its underwriting, asset selection, and ability to source power-secured projects the key near-term proof points; a large addressable-market estimate does not resolve those execution risks. Political resistance may also shift the advantage toward established, power-connected hubs, but mature markets are not immune to grid constraints or local opposition. This favors scrutinizing project-level power and permitting status over broad AI-demand narratives.

Over 1–3 months, watch BXDC deployment disclosures, project approvals, financing costs, and any evidence that tenant commitments are being delayed. Over 6–18 months, constrained power and permitting could make existing facilities more valuable relative to greenfield pipelines, while also limiting sector-wide growth. Higher yields add refinancing and valuation pressure. Oracle’s notice is a project-specific warning about schedule and contractual exposure—not evidence that all Blue Owl assets or leases are impaired. A sharp deterioration in hyperscaler spending would be a separate demand-side risk.

Contrarian angle: political opposition may slow marginal projects without materially displacing demand from existing, well-sited facilities. That supports selectivity, not a blanket short of data-center REITs. Prefer established operators over newly assembled vehicles until execution is demonstrated; validate valuations, leverage, and power availability before sizing.

Key Decisions for Investors

• Relative-value watch/trade: consider a small long DLR/EQIX versus short BXDC only after checking valuation, borrow, liquidity, and portfolio beta. The thesis is operating-asset exposure over unproven deployment; avoid treating either incumbent as insulated from power, rate, or permitting risk. Falsify if BXDC demonstrates timely, attractively underwritten deployment and the relative discount fails to close.
• Keep BXDC on an event-driven watchlist rather than buying the market-size narrative. Require evidence on capital deployment, asset yields, tenant concentration, power contracts, and permitting before upgrading the view; do not assume the trust’s risk profile is equivalent to Blackstone’s consolidated business.
• Treat the Oracle/Project Jupiter issue as a monitor for project-level delay and contract economics, not an automatic short in ORCL or OWL. Reassess if further tenant notices, schedule slippage, or approval setbacks emerge; the thesis weakens if the project reaches operation on schedule and payment terms remain intact.
• For the broader sector, favor existing facilities with verified power access over speculative development exposure. Revisit the cautious stance if permitting accelerates, grid capacity becomes demonstrably available, and hyperscaler demand commitments remain firm despite AI-spending revisions.