Equinix points to smaller facilities as scrutiny of U.S. data center construction grows

Equinix says its data centers, typically 30–60 megawatts, may be less exposed to U.S. proposals generally targeting facilities around 100 megawatts. The company increased capital spending to $7 billion from about $2 billion a few years ago as demand grew; annualized gross bookings topped $500 million in Q3, up 35% year over year. BMI estimates 53% of projected U.S. hyperscaler pipeline growth is in states where newly elected governors could restrict construction, with effects likely more apparent from 2028; analysts also cite power, grid and equipment constraints in Asia.

Analysis

The key distinction is regulatory exposure by facility size versus exposure by total campus load, power demand, or cumulative local impact. Equinix’s stated sub-100MW footprint could be an advantage only if rules are written and enforced at the individual-facility level; aggregation or utility-connection limits could erase it. Treat management’s characterization of proposed thresholds as a company claim, not a confirmed safe harbor.

If restrictions bind large greenfield campuses, Equinix could capture incremental enterprise and cloud demand through colocation and interconnection rather than direct hyperscaler builds. That is a plausible relative advantage over larger-campus-oriented developers such as Digital Realty, but not necessarily incremental industry demand: permitting delays may defer deployments and weaken orders for power, cooling, networking and semiconductor suppliers. The reported 2028 timing makes this a medium-term pipeline risk, not a near-term demand shock.

Asia offers a potential outlet, but does not remove the constraint: power, grid access and local approvals may determine which projects proceed. Equinix’s higher capex raises the importance of utilization, bookings conversion and returns on newly delivered capacity; gross bookings alone do not establish realized revenue or returns. The contrarian risk is that investors capitalize a regulatory advantage before bill language, project-level applicability and power availability are known.

Key Decisions for Investors

• No event-driven trade on the headline alone. Monitor U.S. bill and permitting language for whether thresholds apply per facility, campus, or aggregate power load; the thesis weakens materially if smaller facilities are captured.
• Consider a conditional relative-value long EQIX versus Digital Realty only if rules demonstrably disadvantage large campuses while preserving smaller colocation sites, and EQIX bookings convert into revenue without deterioration in utilization or returns. Reassess on regulatory text and the next earnings update; do not assume the advantage from management commentary alone.
• Track EQIX capex alongside delivered capacity, utilization, bookings conversion and management’s return metrics. A slowdown in bookings conversion, weaker utilization, or evidence that power constraints delay EQIX openings would falsify the operating-upside thesis.
• Keep semiconductor and data-center equipment exposure as a 1–3 month watch item rather than a direct short: the article points to potential project delays, with the larger pipeline effect described as emerging from 2028. Revisit if permitting delays become measurable in hyperscaler project schedules or supplier guidance.