States Put Data Centers on Pause: Analyst Reaction (Podcast)

America’s data center boom is facing political headwinds: Texas, New York, and Pennsylvania are moving to slow or pause new projects over electricity costs and grid/infrastructure concerns. Texas alone represents ~20% of the US pipeline, so even short delays could cost developers billions and raise pipeline risk. The bipartisan pushback adds near-term uncertainty for developers trying to secure power and bring capacity online.

Analysis

This is less a demand shock than a capital-allocation shock: if permitting/local opposition becomes a recurring gate, the market should stop paying up for “certainty of build” and start valuing operators with existing energized capacity and contracted revenue. That is a relative negative for greenfield data-center infrastructure suppliers with backlog tied to new starts, while established colo owners can gain pricing power as hyperscalers substitute leasing for self-builds. The second-order effect is that utility load-growth narratives get pushed right, which can compress multiples for names whose bull case depends on 2-3 years of AI-driven rate-base expansion.

The immediate move is sentiment-driven, but the real catalyst window is 1-3 months: watch whether the pause language spreads beyond a few states, whether interconnection queues start seeing withdrawals, and whether customer capex guides flatten. Over 6-18 months, the risk is a broader re-rating of the entire “AI power” complex if regulators force more local hearings, raise cost burdens, or slow transmission approvals. The thesis is falsified if hyperscalers keep lifting capex and utilities continue approving large-load interconnects despite the political noise.

Consensus is probably underestimating substitution: delays in Texas/Pennsylvania don’t kill demand, they reroute it toward incumbent colo and away from pure-build exposure. That argues for relative shorts in equipment-heavy names and caution on merchant-power enthusiasm if load assumptions get deferred. The overdone part may be any immediate selloff in quality owners with existing occupancy, because scarcity can improve bargaining power even while new project starts pause.

Key Decisions for Investors

• Pair trade: long DLR / short VRT for 1-3 months; the thesis is that paused greenfield builds shift demand toward leased capacity and away from data-center equipment spend. Target ~2:1 upside/downside if permitting headwinds broaden.
• Alternative pair: long EQIX / short ETN on rallies; EQIX benefits from capacity scarcity, while ETN is more exposed to timing risk in power-distribution orders tied to new projects. Exit if ETN guides backlog growth unchanged for two consecutive quarters.
• If the market overreacts, buy 3-6 month VRT put spreads instead of outright short stock; use this only if management commentary starts implying project delays flow into 2027 revenue. Falsify if backlog conversion stays on schedule.
• Keep CEG/merchant-power names on alert, not as immediate shorts: they only become attractive if forward load expectations roll over and power curves soften. Watch for 1-3 month revisions to AI load assumptions before taking action.