Market inflection: capacity follows power and contracts

Investment and build activity have pivoted away from traditional clusters, with a substantial portion of the current expansion now taking place in alternative U.S. regions rather than legacy hubs.

A sizable construction pipeline — roughly 35 gigawatts — is underway, and about 64% of that capacity is being developed outside established markets, signaling a geographic redistribution of supply.

Market tightness persisted through the close of 2025: effective vacancy hovered around 1%, maintained for a second consecutive year, which limits the near-term risk of oversupply.

That tightness is reinforced by demand from large cloud providers and AI compute buyers — the hyperscalers — who have committed to extensive infrastructure spending, and who are pre-leasing the vast majority of new space.

Roughly 92% of facilities under construction include tenant commitments, which compresses speculative exposure and shifts commercial risk toward long-term occupiers and their capital plans.

Capital markets have noticed: lenders and institutional investors funneled record sums into the sector last year, while the top hyperscalers announced programmatic capital plans that will sustain demand into 2026.

Developers are adapting tactics; some are embracing shorter hold periods and a build-and-sell approach to lock returns now while avoiding longer-term operational uncertainty.

Yet the expansion faces a technical and regulatory ceiling: access to reliable grid connections and incremental power capacity remains a bottleneck, with interconnection schedules often stretching multiple years and forcing operators to weigh onsite generation options.

That constraint is a primary reason why regions with available power capacity — notably parts of Texas — are attracting larger shares of new projects and are on track to overtake historic leaders.

A recent regulatory trend in Texas and several other states has intensified the execution risk: regulators and grid operators are reassessing approvals for large data‑center interconnections, conditioning energization on defined upgrades, studies or operational limits. Industry trackers attribute roughly $64 billion of U.S. planned data‑center projects to delays, cancellations or rework tied to permitting and community pushback, a reminder that permits once considered routine can now trigger supplemental technical review.

To reduce community and system impacts, developers are negotiating bespoke mitigations — building substations, accepting connection fees, committing to demand‑response programs or adding on‑site generation and storage. These remedies can lower peak-driven transmission needs but raise capital costs and complicate project timelines.

Concurrently, hyperscalers are changing how they secure power: several leading cloud companies are acquiring operating clean‑energy and development portfolios and prioritizing solar paired with battery storage to meet 24/7 load profiles. That vertical integration shortens delivery timelines and de-risks power supply for large compute customers, increasing the strategic value of operating renewable-plus‑BESS assets and making certain developers acquisition targets.

For investors and occupiers, the combined technical, regulatory and procurement shifts reduce near‑term vacancy risk but heighten execution and financing uncertainty — idle or partially energized capacity ties up capital and shifts bargaining leverage in commercial power negotiations.

Over the next 6–12 months, expect the market to bifurcate: capital-rich buyers and hyperscalers will secure long lead‑time capacity and integrated energy solutions in power‑abundant locales, while developers lacking grid access or energy capabilities will pivot strategies, accept heavier infrastructure obligations, or sell projects to balance‑sheet owners. The outcome will be accelerated consolidation, re‑pricing of power‑ready assets, and clearer delineation of who bears upgrade and delivery risk.