What If the Next Big AI Opportunity Isn’t in Technology?
Just imagine for a moment that one of the fastest-growing technologies in the world depends on something technology cannot create virtually: land and power.
We hear about Artificial Intelligence almost every day. We talk about ChatGPT, automation, technology, and how AI may change the way we work, operate businesses, and live our lives. But what if one of the biggest stories developing around AI isn’t only about technology? What if it’s also a commercial real estate story? AI needs data centers, and data centers need land, enormous amounts of electricity, fiber, infrastructure, and physical real estate. As demand continues to grow, perhaps the bigger question for commercial real estate investors isn’t simply, “How do I invest in data centers?” but rather, “What real estate could benefit from everything being built around them?”
The latest numbers are difficult to ignore. According to CBRE’s North America Data Center Trends report released August 27, U.S. data center vacancy across primary markets fell to a record-low 1.4% during the first half of 2026, even as supply increased 33.7% year over year to a record 10,903 megawatts. Net absorption increased 11.7% year over year to 1,456.2 MW, driven largely by hyperscale and AI users competing for increasingly limited blocks of available power. At the same time, North American data center construction increased 24.8% to 7,481 MW, with more than 80% of capacity under construction already preleased. Think about that for a moment. A tremendous amount of new capacity is being built, yet demand is absorbing it so quickly that vacancy continues to sit at historic lows. (CBRE)
For those of us in commercial real estate, perhaps one of the most interesting parts of this story is how AI may change what we traditionally consider a great location. We have always heard “location, location, location.” But what makes a great location when the end user requires an extraordinary amount of electricity? A property could have hundreds of acres, highway access, favorable zoning, and what appears to be tremendous development potential, but what if sufficient power cannot actually be delivered to the site? Suddenly, access to electrical infrastructure and fiber can become just as important as many of the traditional characteristics CRE investors have considered for decades. JLL reports that 77% of North American data center capacity currently under construction is now located in frontier markets, with areas such as West Texas, Ohio, Louisiana, Indiana, and the Carolinas benefiting from the shift. (JLL) What if some of tomorrow’s most valuable commercial land is sitting in markets investors historically overlooked?
This is where I believe the conversation becomes even more interesting for traditional commercial real estate investors. Not everyone is going to own or develop a massive data center, and perhaps they don’t need to. Potential beneficiaries may include industrial and logistics properties serving construction and operating needs, land with verified access to utilities and transportation, hotels and short-term lodging supporting contractors and visiting personnel, and selected retail or service businesses that could benefit from new employment and population growth. Construction firms, equipment suppliers, engineering companies, and other contractors may also create demand for nearby commercial space. But these are potential beneficiaries, not automatic winners. The strength of the opportunity depends on whether a project is actually funded, whether construction is moving forward, how many permanent jobs it will create, and whether the surrounding market has the infrastructure and demand to support additional development. The relevant question is not simply what properties are nearby, but which properties have a credible connection to measurable activity generated by the project.
That distinction is important because supporting-property opportunities can easily become speculative. A hotel may benefit from construction workers, but that demand could be temporary and may not justify a long-term acquisition price. Industrial space may appear well positioned, but its value depends on tenant requirements, building functionality, access, and competing supply. Land near a proposed data center may attract attention, but proximity alone does not establish development potential. Before treating any property as a beneficiary, investors should underwrite the fundamentals: confirmed utility capacity, documented project milestones, zoning and entitlement status, transportation access, tenant demand, achievable rents, construction costs, absorption, operating expenses, and exit assumptions. They should also test the investment without assuming that every announced project will be completed on schedule or that every projected job and ancillary business will materialize. A real opportunity should be supported by evidence that can be verified; a speculative assumption is a future benefit that has not yet been demonstrated.
Of course, this is also where we need to separate excitement from sound investment strategy. Strong demand does not automatically make every data-center-related property a good investment. Power availability is already becoming one of the industry’s largest challenges. PJM, which operates the largest U.S. electricity market and serves a significant portion of the Mid-Atlantic and Midwest, recently saw its 2028–29 capacity auction fall 6.8 GW short of its reliability requirement, while prices reached the auction’s maximum cap of $325 per megawatt-day. Rapidly increasing electricity demand from data centers is one of the pressures facing the grid, alongside challenges in bringing new generation online. (Reuters) PJM has even proposed a framework that could require certain data centers to rely on backup generation when electricity supplies become dangerously tight. (Reuters) That means investors and developers need to consider much more than acreage and zoning. Where is the power coming from? Is utility capacity actually committed or merely anticipated? What infrastructure still needs to be built? Who is responsible for paying for it? How long could approvals take? What happens to the investment if the data center is delayed, downsized, relocated, or never built? These are underwriting questions, not minor details.
That brings me back to something I believe applies to almost every commercial real estate investment. Following a trend and understanding an opportunity are two very different things. AI and data centers may be one of the most exciting growth stories in commercial real estate today, but that does not mean we should automatically chase anything associated with them. Perhaps we should be asking better questions. What is driving demand in that particular market? What infrastructure already exists? What future supply is planned? Who are the likely users? Which benefits are supported by signed leases, utility commitments, permits, financing, or construction activity? Which benefits are merely projections? And most importantly, does the investment still make financial sense if the project takes longer, creates fewer jobs, requires more infrastructure spending, or generates less surrounding demand than expected?
Commercial real estate has always evolved alongside changes in the way we live and do business. Railroads influenced where cities developed. Automobiles reshaped suburbs and retail. E-commerce transformed warehouses and logistics. Remote work changed how we think about office space. What if Artificial Intelligence is beginning the next major transformation? Perhaps the opportunity isn’t simply identifying the next popular asset class. Maybe it is understanding what that transformation requires, where it needs to happen, and which properties have a defensible, measurable connection to the resulting demand. AI may be a technology revolution, but increasingly, it is also a land, power, infrastructure, and commercial real estate story.
And perhaps we are only beginning to understand what that could mean.



