In AI Real Estate, Power Is the New Location

At first, the data center story was easy to follow. AI demand increased, and more facilities were needed to support it. That logic still holds, but it no longer explains the full picture.

The pressure has shifted into something more basic. It is not just about building space or land availability. It is about whether the site can actually be powered at scale, and whether that power can be delivered on time.

Without that, the rest of the project does not matter.

When power becomes the limiting factor, the way assets are judged starts to change.

A well-located site can lose its edge if it cannot secure enough electricity or if the timeline stretches too far. At the same time, less obvious locations can gain value if they have a clear and reliable path to power delivery. This flips the usual logic. Location still matters, but it is no longer the only priority. The ability to operate becomes just as important as where the property sits.

According to JLL's 2026 Global Data Center Outlook, the average wait time for a grid connection in primary markets now exceeds four years, while construction costs have risen from $7.7M to $10.7M per MW between 2020 and 2025 — a 7% CAGR — with a further increase to $11.3M per MW forecast for 2026. 

Roughly 100 GW of new capacity is expected to come online by 2030, requiring up to $3 trillion in total investment, and AI workloads could represent half of all data center demand by that point, up from approximately a quarter in 2025.

Developers are now evaluating more layers at once. It is not just land size or connectivity. It is capacity, timing, and flexibility.

They want sites that can handle large loads and expand over time. They also need those sites to move quickly through the approval and build process. Delays in power delivery can stall the entire project, even if everything else is ready.

This is starting to reshape which markets attract investment and how land gets priced within them.

As projects grow larger, they also draw more attention. Communities are paying closer attention to how these facilities affect local resources.

Questions around power usage, infrastructure strain, and long-term impact are becoming more common. This is not a minor hurdle. It is part of the deal.

If a project loses local support, it can slow down or stop altogether. Developers now have to manage not just construction and financing, but also how the project is received at the community level.

Securing land early is no longer enough. The real challenge is moving through each stage without friction.

Power agreements, infrastructure upgrades, and local approvals all have to align. If one part breaks down, the entire project can weaken.

This creates a wider gap between strong and weak opportunities. Projects that look similar at the start can end up very different once execution begins.

This shift is not limited to one asset class. It reflects a broader change in how real estate works when infrastructure becomes tight.

Value moves toward assets that can actually function within real-world limits. That includes power, timing, and the ability to maintain support over the life of the project.

In that environment, the best assets are not always the most obvious ones. They are the ones that can move from plan to operation without breaking.

In the AI buildout, the building is only one part of the equation.

Power access, delivery timing, and local acceptance now define whether a project succeeds. The sites that win are the ones that can carry all three without delay.