Northern Virginia has achieved what every market wants: near-zero vacancy and massive pre-leasing. The problem? There's nowhere left to build. With just one percent vacancy maintained for 12 consecutive quarters and 87 percent of existing construction already pre-leased, the market has become a victim of its own success.

I moderated the "Built for Speed" panel on solving Virginia's data center construction crunch at DCD>Connect | Virginia 2025 in November, and the irony wasn't lost on anyone. We're sitting in the market that launched the modern data center industry, talking about where else to build next.

The numbers tell the story. Nationally, colocation vacancy sits at 2.3 percent, with 73 percent of the 8GW in our development pipeline already spoken for. A market running at greater than 98 percent utilization is either impressive or terrifying – impressive if you're watching from the sidelines, but terrifying if you're trying to find space.

The shift to frontier markets

Mario Sawaya, senior vice president, global head of data centers and technology at Aecom, made the shift clear during our discussion. Sawaya's firm operates data center projects in 40 countries, and the executive is seeing clients completely pivot their site selection strategy. They're asking about placing small modular reactors (SMRs) on sites in non-attainment zones four years from now.

That's not theoretical planning. That's business reality when traditional markets can't accommodate demand growing at 24 percent annually.

The conversation becomes more interesting when you consider what's driving developers to these frontier markets. It's not just about finding available land. Vincent Calvo, director of data center infrastructure build at Oracle, confirmed what we're seeing across the industry. The most successful deals Calvo's team closes are with providers who secured land and power before everyone else arrived.

Translation: The winners aren't necessarily building better data centers. They're building in better locations before those locations become obvious to everyone else.

Follow power, not patterns

This shift creates fascinating dynamics around construction and development. We're seeing constraints around power delivery, zoning, NIMBY (Not in My Backyard) opposition, and permitting drive innovation in unexpected directions. Some developers are future-proofing campuses for SMR technology that won't come online until 2030 or 2031. Others are exploring partnerships with federal agencies on government land (despite well-founded concerns about bureaucratic timelines).

The geographic expansion isn't random. Developers are following a simple formula: find the intersection of available power and developable land. Sometimes that means non-attainment zones. Sometimes it means rural markets with transmission infrastructure but no existing demand. Sometimes it means brownfield sites with existing electrical capacity.

Why this time is different

What's particularly interesting is how this expansion challenges traditional risk models. Karen Petersburg from American Real Estate Partners (AREP), who has 8GW in the pipeline, raised a critical point during our panel about the timing mismatch between planning cycles and customer signals. How do you plan four or five years out when your customers' demand signals arrive with 12 months' notice or less?

The answer seems to be accepting that the old playbook doesn't work anymore. The developers winning contracts aren't necessarily the ones with the best engineering or the lowest costs. They're the ones who anticipated demand patterns and secured sites before those patterns became obvious to competitors.

This creates both opportunity and risk for builders and developers. The opportunity is clear: get to frontier markets early and secure outsized returns. The risk is equally clear: guess wrong about which markets will develop, and you're holding expensive dirt with no revenue potential.

Vince Calvo, director of data center delivery at Oracle, agrees, noting that the key issue right now is “100 percent power” and that early planning is essential so that customers can be presented with a turnkey powered base build. Oracle, like other Hyperscale leaders, is embracing change to address power constraints, exploring gas turbines, fuel cells, and other new innovative power delivery technologies and fuels.

Nancy Novak, chief innovation officer at Compass Datacenters, shared the broader context of how important cross-industry collaboration is to surmounting delivery challenges. Compass has created deep partnerships across its supply chain, thereby buffering inventory to ensure delivery continuity across years, not months. There are other opportunities, noted Novak, such as modern contractor models that incentivize on-time product delivery.

But here's what makes this moment different from previous expansion cycles: The demand drivers aren't cyclical. Global data growing at 24 percent annually shows no signs of slowing, and nothing suggests that trajectory will change. The applications driving this growth (e.g., artificial intelligence workloads, Edge computing, autonomous systems) aren't fashion trends that might fade.

That means the current scramble for sites isn't a temporary phenomenon – it's the new baseline. The developers who figure out how to identify and develop frontier markets won't just survive the current construction crunch. They'll define the next generation of data center geography.

But there’s also some hope for existing, highly developed markets like NoVa; retrofitting, or what JLL calls “FutureFitting”. With such low colocation supply, escalating demand, constrained power, and- now- Wall Street’s urgent demand for reduced time-to-revenue, requests to remodel standing data halls to support 50kW+ cabinet densities by implementing liquid cooling are skyrocketing. At roughly half the cost of a greenfield build, significantly less time to revenue, and the sustainability benefits of circular reuse, it’s no wonder that the US data center renovation is on track to hit $9B in 2025, growing at an 18 percent CAGR (source: Global Market Insights US Data Center Renovation Market, November 2025).

The question isn't whether to expand beyond traditional markets or invest in existing. It's whether how you’ll strategically plan and collaborate to serve customers in both.

Ready to move beyond saturated markets? Want to run through retrofit scenarios to see what investment offers best risk and return profiles? Partner with the experts who know where to look next. Contact JLL's data center specialists to explore opportunities in frontier markets ahead of your competitors.