QTS Co-CEO on Building Eco-Friendly Data Centers

QTS Co-CEO Tag Greason said the AI boom is fundamentally an infrastructure buildout, but expanding data centers faces growing political and environmental backlash. He argued that U.S. AI leadership will depend on maintaining leadership in the underlying data-center and power infrastructure, highlighting permitting, community, and environmental risks to sector expansion.

Analysis

The binding constraint is shifting from server availability to power interconnection, which changes the AI value chain’s economics. Permitting delays and local opposition raise the scarcity value of data-center campuses with contracted power, but also increase development carrying costs and make speculative capacity less attractive. Listed incumbents DLR and EQIX should retain pricing power in supply-constrained Tier-1 markets, while smaller developers without secured transmission rights face materially higher execution risk.

The cleaner expression is likely upstream: VRT, ETN, PWR and GEV benefit whether projects proceed now or are redesigned around distributed generation, substation upgrades, backup power and higher-density cooling. A delayed campus does not necessarily eliminate equipment demand; it can increase electrical infrastructure content per megawatt as operators pursue on-site generation and grid-hardening. Over 6-18 months, utilities with excess generation and favorable regulatory frameworks—CEG, VST and select regulated utilities—can monetize long-duration contracts, although regulators may seek to socialize grid costs and cap returns.

Near term, political headlines are more likely to pressure data-center REIT multiples than impair already-contracted revenue. The contrarian point is that opposition is not uniformly bearish for the sector: restricted new supply can support rent and renewal spreads for operational facilities. The thesis fails if hyperscalers materially slow lease commitments, if power-delivery lead times improve faster than expected, or if state regulators impose data-center-specific connection fees that make signed projects uneconomic.

Watch 1-3 month indicators: utility interconnection queues, announced hyperscaler pre-leases, DLR/EQIX leasing spreads, and transformer/switchgear lead-time commentary from ETN and VRT. A broad slowdown in pre-leasing—not isolated project opposition—would be the signal that AI infrastructure demand is becoming demand-, rather than supply-, constrained.

Key Decisions for Investors

• Prefer a 6-12 month long VRT/ETN basket over DLR/EQIX: electrical-content intensity rises under grid constraints, while REITs absorb permitting and construction-duration risk. Use a relative-value structure rather than outright exposure; exit if VRT/ETN order growth or backlog conversion weakens for two consecutive reporting periods.
• Accumulate PWR on permitting-related weakness over 3-9 months: incremental transmission, substation and interconnection work is a second-order beneficiary of constrained grid capacity. Key risk is utility capex deferral; reassess if major utility capital plans shift from transmission expansion toward ratepayer affordability.
• Maintain a watch, not a recommendation, on CEG and VST for 6-18 month power-contract upside. Initiate only upon independently disclosed incremental long-term data-center load contracts and visibility into price/volume terms; regulatory treatment of grid-upgrade costs is the critical missing variable.
• Avoid chasing a blanket short in data-center REITs after adverse local headlines. Consider reducing DLR/EQIX only if pre-leasing decelerates alongside lower guidance for development starts; isolated permitting denials may instead tighten supply and support existing-asset valuations.