AI Data Center Land Purchases Surge 79% Amid Rising Power Premiums

Executive Summary

U.S. land purchases for AI data centers hit $6 billion in the first half of 2023, a 79% year-over-year increase, as developers raced to secure sites with dependable power supply. Power premiums — the cost differential commanded by land with confirmed, reliable grid access — have become the dominant variable in site selection, eclipsing traditional real estate factors like location and entitlement status. Developers who control sites with firm interconnection have the clearest path to monetization; those chasing underpowered land face compounding cost and timeline risk. The InfraSale takeaway: power access is no longer a secondary diligence item — it is the primary acquisition criterion.

What Happened

U.S. data center land purchases for AI-related development reached $6 billion in the first half of 2023. That figure represents a 79% increase compared to the same period in 2022 — a growth rate that signals structural demand, not a one-cycle spike. The driver is straightforward: AI workloads require significantly more power per square foot than traditional enterprise computing, forcing developers to compete aggressively for sites where utilities can actually deliver capacity at scale.

Developers across the market are actively prioritizing sites with confirmed power availability, in some cases paying substantial premiums over comparable land that lacks grid certainty. The result is a bifurcated market: powered land is appreciating sharply, while unpowered parcels are increasingly difficult to move.

Source: TechTarget

Why This Matters

A 79% year-over-year jump in a single asset class is a structural signal, not a headline anomaly. The AI infrastructure buildout — driven by hyperscalers and large colocation operators — is compressing the development pipeline in a way that conventional data center cycles never did. Sites that would have been considered speculative two years ago are now commanding acquisition prices that reflect near-term certainty of tenant demand.

The power premium dynamic is self-reinforcing. As more capital chases fewer grid-ready sites, those sites appreciate faster, which raises the cost basis for all subsequent development. Developers who lock in power agreements now are effectively acquiring a competitive moat; those who don't face cost escalation at every stage of the development stack.

This also changes the risk calculus for landowners. Raw land adjacent to existing substations or transmission infrastructure is suddenly a strategic asset, not a passive holding. The first-mover advantage for landowners who understand this shift is measurable and time-limited.

Power & Interconnection Impact

Power availability has moved from a permitting checkbox to the lead variable in underwriting. Developers are scrutinizing substation proximity, available transformer capacity, and utility interconnection queue position before making offers — criteria that previously entered the diligence process much later. In many markets, the interconnection queue itself has become a de facto barrier to entry, with new applicants facing multi-year wait times before capacity is confirmed.

Industry context: Regions with deregulated wholesale power markets — ERCOT, PJM, MISO — have historically offered more flexible procurement structures. That flexibility is now a primary site-selection filter as developers attempt to lock in competitive power costs before load growth erodes available headroom.

Increased demand for grid capacity at scale will pressure local utilities to accelerate transmission upgrades. Where utilities move slowly, developers will either pay premium rates for existing capacity or absorb the cost of substation improvements — both of which reduce project returns.

Land, Zoning & Permitting Impact

Zoning codes in most jurisdictions were not written with hyperscale data center campuses in mind. As developers target sites at scale — often 500 to 2,000+ acres to accommodate campus buildout and associated cooling infrastructure — they are increasingly running into agricultural, industrial, and mixed-use zoning designations that require rezoning or conditional use approvals.

Permitting timelines are being compressed on the demand side while regulators lack the staff and precedent to move faster. The result is schedule risk. A project that closes land in Q1 may not receive a building permit until Q3 or Q4 of the following year, depending on the jurisdiction — and interconnection approval timelines are entirely independent of local permitting clocks.

Local governments in emerging data center markets are beginning to adapt, in some cases creating data center overlay zones or fast-track permitting pathways. Assumption: communities with existing industrial infrastructure and established utility relationships are more likely to offer streamlined entitlement processes, giving those markets a near-term edge in attracting capital.

Investment Takeaway

InfraSale Marketplace

Turn this intelligence into a deal

InfraSale connects landowners, developers, and tenants directly — skip the broker chain.

The $6 billion figure recalibrates how investors should think about powered land as an asset class. Specific implications:

Power-confirmed land commands a structural premium. Sites with utility commitments or existing substation interconnection should be underwritten at a separate tier from unentitled land, regardless of comparable sales in the area.
Unpowered land adjacent to transmission infrastructure is the next bid layer. As Tier 1 powered sites tighten, capital will move to sites that can achieve grid access within 18–36 months.
Zoning risk is repriced upward. Projects requiring discretionary approval — rezonings, conditional use permits, environmental impact review — now carry measurable schedule risk that affects IRR assumptions.
Local utility relationships are balance sheet items. Developers with established utility partnerships can access capacity before it hits the open market. That relationship has tangible value.
Competing asset classes feel pressure. Industrial, logistics, and other power-hungry land uses are now competing with data center developers for the same grid-ready inventory, pushing prices across all categories.

InfraSale Market Angle

For developers actively sourcing AI data center sites, the $6 billion figure is a benchmark and a warning. The market for grid-ready land is tightening in every major metro and emerging secondary market simultaneously. Waiting for price discovery in a new submarket before moving is increasingly costly — by the time comparable sales confirm the premium, the best sites are under LOI.

Developers who win in this environment share one characteristic: they have built systems for identifying powered land before it reaches the open market. That means direct relationships with utilities, landowners near substations, and platforms that surface interconnection-ready parcels before they are formally listed.

For investors capital-allocating into this theme, the underlying question is not whether AI infrastructure demand is real — it demonstrably is — but whether the site you are underwriting has the power position to support the projected development timeline. A site without power certainty is not a data center asset; it is an option on a data center asset.

Market Signal

Location: Unspecified
Primary Issue: Rising power premiums
Infrastructure Theme: Land acquisition competition
Who Benefits: Developers with access to sites with reliable power
Who's at Risk: Developers facing increased land costs and competition
InfraSale Takeaway: Developers should prioritize sites with stable power access to remain competitive.

Take Action

The window for securing grid-ready land at pre-premium pricing is closing in most active markets. Developers and investors who move now — with a disciplined focus on power confirmation and zoning clarity — will hold a structural cost advantage over those who enter the market in 12 to 18 months. Browse available powered land and DC sites.

FAQ

What are the primary factors driving land acquisition for AI data centers?

Power availability is the leading driver. AI workloads require substantially more energy per rack than conventional computing, making grid access and substation proximity the critical site-selection variables. Competition for a limited inventory of powered sites is pushing acquisition prices higher across every major market.

How do rising power costs affect data center investments?

Power premiums compress operating margins and increase the cost basis of land acquisition, which directly affects projected returns. Developers underwriting sites without confirmed power agreements are effectively holding an open variable in their financial model — one that can materially shift IRR if utility commitments are delayed or priced above projection.

What zoning challenges might developers face in acquiring land?

Many high-potential data center sites sit in agricultural or light industrial zones that require rezoning or conditional use approval before construction can begin. Those processes are discretionary, community-facing, and time-variable — introducing schedule risk that is difficult to hedge and increasingly priced into acquisition negotiations.

How should developers evaluate interconnection queue position when acquiring land?

Interconnection queue position is independent of local permitting and can extend project timelines by two to four years in constrained markets. Industry context: developers are increasingly paying to assume existing queue positions — or acquiring utility-side assets — to bypass standard queue timelines and de-risk project schedules.

Is the $6 billion figure a one-time spike or a durable trend?

A 79% year-over-year increase over a six-month measurement period, driven by structural demand from AI compute buildout, is consistent with a durable trend rather than a single-quarter anomaly. Assumption: absent a significant slowdown in AI infrastructure investment by hyperscalers, land acquisition volumes are likely to remain elevated through at least 2025.