Almost Half of US Data Centers That Were Supposed to Open This Year Slated to Be Canceled or Delayed

About half of U.S. data centers slated to open in 2026 face delays or cancellations; of 12 GW planned to open this year, only ~33% are actually under construction. For 2027, just 6.3 GW are under construction versus 21.5 GW announced, and 37 GW have no firm completion date with only 4.5 GW started. Causes are supply-chain bottlenecks in critical electrical components (batteries, transformers, breakers) sourced internationally, lengthening build times and threatening capacity rollouts for AI workloads.

Analysis

The choke point is not demand for computing but the midstream electrical hardware and permitting ecosystem that stitches power to servers. That means near-term pain will concentrate on project deliverability and cashflow timing for owners and contractors—capex is being rephased into later quarters rather than erased—which compresses near-term revenue for build-heavy players while increasing replacement-value optionality for completed assets. On a 3–24 month view, expect price and lead-time spikes for transformers, breakers, and battery systems; on a 2–5 year view, expect a politically-driven push to onshore critical electrical component production that re-rates manufacturers with domestic capacity or easy scale-up.

Second-order winners aren’t the cloud hyperscalers per se but the domestic electrical OEMs, integrators with stocked inventory, and modular/edge builders who can deploy without protracted utility interconnection. Conversely, wholesale data-center operators with large forward-build portfolios and thin near-term occupancy assumptions will face guidance risk and covenant pressure as projects shift. Local permitting and community pushback amplify these dynamics: projects that can be sited in pre-permitted industrial parks or retrofitted shell space will transact at a premium versus greenfield builds buried in approval cycles.

Catalysts that would reverse the current squeeze include accelerated US/NAFTA manufacturing investments (18–36 months), targeted tariff or export-policy relief that speeds inbound supply (weeks–months), or an unexpected demand pullback for AI workloads that reduces urgent component ordering (quarters). Tail risks: a geopolitically-triggered shortage of specific electronic-grade components or a credit squeeze that forces cancellation of projects en masse. Monitor inventory build at major OEMs, permit-issuance trends in top metro nodes, and guidance from REITs/hyperscalers for early-warning signals.

Key Decisions for Investors

• Pair trade (3–12 months): Long Eaton (ETN) equity, Short Digital Realty (DLR) — rationale: ETN benefits from higher prices and accelerated onshoring of switchgear/transformer demand; DLR faces lease-up and construction cadence risk. Position size: 2:1 notional in favor of ETN. Risk/reward: asymmetric — ETN upside if onshoring political momentum or inventory tightness persists; downside if macro capex collapses. Use 12% stop on ETN leg, 15% stop on short DLR leg.
• Options play (12–24 months): Buy ABB Ltd (ABB) 2027 LEAP calls funded by selling nearer-term vertical spreads — rationale: capture multi-year re-rating as European/Swiss OEMs win North American content share and aftermarket spare-parts pricing normalizes. Target: 2.5x payoff if onshoring accelerates; max loss = premium paid. Reduce position if announced NA capacity expansion timelines slip beyond 36 months.
• Short REIT knee-jerk (3–9 months): Short Equinix (EQIX) or use OTM put calendar spreads to monetize expected downside volatility around quarterly guidance cycles — rationale: occupancy/FFO misses likely if build pipelines slip. Risk management: size to <2% NAV, set a 10–15% hedge-trigger to pare exposure if broader market rallies.
• Tactical commodity/industrial hedge (6–18 months): Long copper futures or mining/transformer steel suppliers via selective miners or HRC steel exposure; alternatively buy short-dated call options on GE (GE) to capture grid-equipment upside. Rationale: component buildouts increase raw-material intensity; hedge downside if demand softens. Target payoff 1.5–3x, keep tenor <18 months to avoid carry drag.