This article is for informational purposes only and does not constitute financial or investment advice. Always consult a qualified real estate or financial professional before making any investment decisions.
There is a new kind of land rush happening across America. It does not look like the Oklahoma Land Run of 1889, and it bears no resemblance to the California Gold Rush of 1849. There are no covered wagons, no prospectors with pickaxes, no claims staked in dusty riverbeds. This one happens in boardrooms and on satellite maps. It involves private equity firms, hyperscale technology companies, and real estate investors quietly acquiring parcels of land in rural Texas, the Ohio countryside, and the backroads of South Carolina — not because of what sits on the land, but because of what runs beneath or beside it.
Transmission lines. Substations. Grid capacity. Fiber routes.
In 2026, the most valuable real estate in the United States is not necessarily a Manhattan skyscraper or a beachfront acre in Malibu. In some cases, it is a flat, unremarkable parcel of land in a secondary market — one that happens to sit within reach of a high-voltage power line and a completed infrastructure study. That parcel might sell for $800,000 per acre. A comparable parcel two miles away, without those characteristics, might sell for $5,000 per acre.
The data center land rush is real, it is reshaping where value gets created in American real estate — and in 2026, it is getting significantly more complicated.
What “Power-Ready” Land Actually Means — and Why It Now Defines Site Selection
To understand why land values near power infrastructure have exploded, it helps to understand what modern data centers actually require — and how dramatically those requirements have shifted in recent years.
A data center is, at its most basic level, a building full of servers. But the servers powering today’s artificial intelligence workloads — training large language models, running inference at scale, processing the compute demands of hyperscale cloud platforms — are categorically more power-intensive than anything that came before them. The physical density of power consumption in next-generation AI data centers, measured in kilowatts per rack, is multiples of what facilities built a decade ago were designed to handle.
This means the foundational question in data center site selection is no longer about square footage, zoning, or proximity to a major city. It is about one thing above all others: where is the power?
“Power-ready” or “AI-ready” land refers to parcels that already have the infrastructure prerequisites that allow a data center to be built and energized on a reasonable timeline. That typically means proximity to high-voltage transmission lines, an existing or near-ready substation with available capacity, access to high-capacity fiber networks, and in many cases a source of cooling water. The most critical of these — and the scarcest — is grid capacity.
Interconnection queues in major U.S. grids currently run five to eight years in many regions. For a technology company trying to deploy AI infrastructure on a compressed timeline, a site that requires a five-year wait for grid interconnection is effectively unusable. A site with a completed power study and available substation capacity is worth an enormous premium — not because of what it is today, but because of how quickly it can become something else.
The Demand Driving All of This: AI, Hyperscale Cloud, and an Electricity Supercycle
The data center land rush is not a speculative phenomenon. It is the direct consequence of a structural shift in the computing requirements of the global economy — one that shows no sign of reversing.
Generative artificial intelligence has fundamentally changed the scale of computation that large technology companies require. Training a single large language model can consume as much electricity as thousands of U.S. households use in a year. Running inference at the scale that major AI platforms operate requires continuous, massive computational power deployed across hundreds of thousands of servers. Every major technology company in the world is racing to expand its AI infrastructure simultaneously.
The aggregate effect on electricity demand is extraordinary. According to a World Resources Institute analysis, the United States has over 100 gigawatts of data center demand coming online between 2024 and 2035 — roughly ten times New York City’s entire summer peak electricity demand.
The numbers are accelerating. The U.S. Energy Information Administration reports that electricity demand has grown approximately 1.7% annually since 2020, largely driven by data center expansion. The EIA forecasts that growth rate will accelerate to 1.9% in 2026 and 2.5% in 2027. In its high-demand scenario, that trajectory would increase reliance on natural gas plants and raise wholesale electricity prices — particularly in Texas through the ERCOT grid and across the Midwest through the PJM interconnection.
A February 2026 study from the Electric Power Research Institute estimated that U.S. data centers consumed between 177 and 192 terawatt-hours of electricity in 2024 — approximately 4 to 5% of national demand. By 2030, EPRI projects that figure could rise to between 380 and 793 terawatt-hours, representing 9 to 17% of total U.S. electricity consumption. The width of that range reflects genuine uncertainty about how quickly AI workloads will scale — but even the low end of the projection represents a transformation of the U.S. electricity system.
The hunt for power-ready land, in other words, is not just a real estate story. It is increasingly entangled with fuel costs, wholesale electricity price volatility, and the structural capacity of a grid that was not built with this demand trajectory in mind.
The Numbers Behind the Land Rush
The price escalation for power-ready land is not subtle. It is one of the most dramatic valuation shifts occurring anywhere in the American real estate market right now.
In Texas, parcels with completed power-infrastructure studies — the documents that confirm a site can receive adequate power delivery and identify the upgrades required — were selling for approximately $20 per square foot, or roughly $800,000 per acre, during 2024 and 2025. Similar parcels in the same geography, without those studies, were trading for $2,500 to $10,000 per acre. The power study, which costs between $50,000 and $300,000 to commission, was translating into a land value premium of 80 to 300 times on an acre-for-acre basis.
In Northern Virginia — home to Ashburn, Virginia, which has earned the nickname “Data Center Alley” and represents the largest concentration of data center capacity anywhere in the world — land has been trading at more than $3.5 million per acre. In 2024, Microsoft acquired a 124-acre parcel in Prince William County for $465.5 million — a transaction that works out to more than $3.75 million per acre for undeveloped land.
In Arizona, a developer paid $136 million for a 2,100-acre parcel in Buckeye — not because the land itself was exceptional, but because the acquisition secured access to power capacity for future data center development. In Texas, landowners who had previously invested in power infrastructure upgrades to support cryptocurrency mining operations are now selling those sites to data center developers at extraordinary premiums. The infrastructure investment made for one technology cycle is proving highly valuable in the next.
Power-ready land commands two to four times the price of standard industrial parcels in comparable locations — because it eliminates the years of waiting, uncertainty, and capital expenditure associated with securing grid access from scratch.
From Square Footage to Megawatts: How the Industry Has Pivoted
The recognition that power capacity — not physical space — is the binding constraint in data center development has triggered a fundamental shift in how real estate firms and investors approach the sector.
Hines, one of the largest real estate investment and development firms in the world, has pivoted from focusing on constructing data center buildings to securing megawatts of power. Hines’ internal research estimates that 40,000 acres of powered land will be needed over the next five years to meet anticipated data center demand — while only approximately 20,000 acres currently exist under operational data centers. That 20,000-acre gap between supply and projected demand is the investment thesis behind an entirely new category of real estate strategy.
In 2025, private equity firm Silver Lake launched a $400 million “powered land platform” — a dedicated investment vehicle focused specifically on acquiring sites with existing grid connections. The logic is straightforward: grid connection rights and completed infrastructure studies have become tradeable assets in their own right, with a value that can be monetized independently of whether a data center is ever built on the site.
This represents a meaningful evolution in how sophisticated capital thinks about real estate. The value is no longer primarily in the physical structure. It is in the invisible infrastructure: the kilowatts of available capacity, the completed study that sits in a utility’s queue, the fiber route that runs adjacent to the property line.
Geographic Shifts: Where the New Data Center Markets Are Forming
The original data center hubs — Northern Virginia and Silicon Valley — are reaching the limits of their power capacity. New projects in these markets face the longest interconnection queues and the highest land prices.
As a result, development activity is migrating to markets that offer a combination of available power, lower land costs, and shorter interconnection timelines.
Texas has emerged as one of the most active secondary markets, driven by its deregulated electricity market, relatively low power costs, and large land supply. The Midwest is attracting increasing attention, with Ohio and Indiana both seeing data center investment. South Carolina has become one of the most prominent Tier-2 data center destinations in the country — Google invested $2.9 billion in Berkeley County and Meta announced an $800 million data center campus in Aiken County.
However, the geographic picture is more complicated than simply identifying where development is moving. It also requires understanding where policy is moving — and in 2026, that story is changing rapidly.
The White House Steps In: The Ratepayer Protection Pledge
In March 2026, the Trump Administration brokered what has become known as the Ratepayer Protection Pledge — an agreement with seven major AI companies including Amazon, Google, Meta, Microsoft, OpenAI, Oracle, and xAI. The pledge requires these companies to build, bring online, or purchase the generation capacity needed to power their data centers, and to pay for all new transmission infrastructure upgrades required to connect them to the grid.
Critically, the pledge stipulates that companies will pay for power capacity whether they use it or not — a take-or-pay structure designed to ensure that the costs of AI infrastructure buildout do not get passed to retail electricity customers.
The policy intent is to protect ordinary consumers from the rate increases that would otherwise result from utilities investing in new generation and transmission capacity primarily to serve hyperscale data center demand. The practical implication for investors and developers is significant: future data center projects will carry higher upfront infrastructure obligations than their predecessors. The cost of securing power access — already elevated by long interconnection queues and infrastructure study expenses — is now more explicitly the responsibility of the technology companies and developers themselves.
This changes the economics of data center development at the margin. Projects that might have been viable under an assumption that transmission costs would be socialized across the rate base now need to underwrite those costs directly. That raises the bar for project feasibility and, paradoxically, further increases the premium on sites that already have power infrastructure in place — because those sites avoid the largest new cost obligations.
The Pipeline Is Slowing — and That Matters
Not all the news from the data center land rush is about acceleration. There are meaningful signs that the pace of development is beginning to moderate — and investors should understand why.
Analysts at Wood Mackenzie note that only 25 gigawatts of new capacity was added to the U.S. data center pipeline in the fourth quarter of 2025 — half the amount added the previous quarter. The slowdown reflects the fundamental constraint that utilities cannot deliver power fast enough to keep pace with developer demand. Many projects currently queued for interconnection may never be built, because the timeline for grid capacity delivery extends beyond what project economics can support.
Some hyperscale companies are now evaluating on-site natural gas generation as a way to bypass long interconnection queues entirely — building behind-the-meter generation capacity that allows a data center to operate independently of the grid for its primary power supply. This approach trades one set of risks — grid interconnection timeline uncertainty — for another, including fuel cost exposure, emissions regulatory risk, and the capital cost of building generation alongside the data center itself.
The cooling of the new project pipeline does not mean the land rush is over. It means the land rush is becoming more selective. Sites that can actually deliver power on a timeline consistent with hyperscale operators’ deployment schedules are becoming more valuable, not less, as the overall supply of developable sites proves more constrained than earlier projections assumed.
State Legislatures Are Responding — and Not Always With Open Arms
Perhaps the most consequential development for the data center land rush in early 2026 is the acceleration of state-level regulatory activity — and much of it is not favorable to unconstrained development.
A February 2026 legislative tracking report documented that more than 300 data center-related bills were introduced across 30 states within the first six weeks of the year alone. The range of legislative responses is wide.
Several states — including New York, South Dakota, and Oklahoma — have proposed moratoriums on new hyperscale data center development while they study the impacts on water supplies, utility rates, and local communities. These are not fringe proposals. They reflect genuine concern from state legislators and utility regulators about the pace of load growth and its consequences for the constituents they represent.
Other states are rethinking the tax incentive structures that were designed to attract data center investment. Many of these incentives were created in an era when data centers were seen as unambiguous economic wins — large capital investments that created jobs and tax revenue without placing significant demands on public services. The current generation of hyperscale AI data centers challenges that calculus. They are more capital-intensive and less labor-intensive than their predecessors, and their electricity demand is large enough to materially affect utility rates for all customers. Several states are introducing special rate classes or demand-response requirements that ensure data center operators pay the full cost of their grid connections — shifting the cost burden that the Ratepayer Protection Pledge attempts to address at the federal level.
For investors underwriting data center land positions, this regulatory environment is a material variable. A site in a state that introduces a moratorium, revises its tax incentive structure, or imposes new environmental review requirements faces a different development timeline and cost structure than the same site would have faced 18 months ago. State-level policy risk is no longer a tail risk in data center real estate underwriting. It is a front-and-center consideration.
Electricity Prices Are Already Rising
The cost consequences of data center demand growth are not theoretical. They are already showing up in electricity bills.
The Sustainable Energy in America Factbook, published by the Business Council for Sustainable Energy, notes that U.S. retail electricity prices rose approximately 2.3% year-over-year, with data center demand expansion cited as a contributing factor. In regions where data center load growth has been most concentrated — Northern Virginia, parts of Texas, the Mid-Atlantic states served by PJM — the rate pressure has been more pronounced.
Rising retail electricity prices create a feedback loop that affects data center economics directly. Higher electricity costs increase the operating expenses of data centers, which are among the most electricity-intensive buildings on earth. They also increase the political salience of data center development among residential and commercial ratepayers who are watching their bills rise. That political dynamic is one of the drivers behind the legislative activity documented above.
For investors focused on the land side of the equation, rising electricity prices are a more ambiguous signal. They increase the value of behind-the-meter generation and the premium on sites that can access lower-cost power sources. They also increase the scrutiny that regulators apply to new large-load interconnection requests. The net effect on land values depends heavily on location and on how utilities and regulators respond in specific markets.
The Risks Are Real and Deserve Honest Assessment
The data center land rush carries genuine risks that deserve direct acknowledgment.
Forecasting electricity demand is genuinely difficult. The EPRI’s 2026 projections for data center electricity consumption span from 380 to 793 terawatt-hours by 2030 — a range so wide that it encompasses scenarios ranging from significant but manageable grid stress to a near-transformation of the U.S. electricity system. The gap between those two outcomes is enormous, and no credible analyst claims high confidence in which trajectory will materialize.
Speculative interconnection requests are distorting grid planning. Developers filing requests to hold a position in a queue without firm development plans create phantom load projections that misallocate grid investment. Utilities and grid operators are working to address this through deposit requirements and more rigorous study processes, but the problem continues to affect the accuracy of pipeline estimates.
Local opposition and environmental constraints can derail projects regardless of their power credentials. A data center rezoning request in Hays County, Texas, was denied in 2025 — after transmission line upgrades had already been completed — because of water consumption concerns. That outcome illustrates a fundamental reality: power availability is necessary but not sufficient. A viable data center site requires buildable land, adequate water supply for cooling, compatible zoning, and community acceptance.
Water is an increasingly material constraint as development expands into arid regions. Parts of Texas, Arizona, and the Mountain West that offer available land and power are also facing long-term water availability pressures. The intersection of data center cooling requirements and regional water stress is a risk that institutional underwriters are incorporating into site feasibility analysis with increasing rigor — and that state legislators are beginning to address through moratorium proposals and environmental review requirements.
The Broader Infrastructure Supercycle
The data center land rush is the most visible current manifestation of a broader structural shift in what kinds of land hold value in the United States — but it is not the only one.
A LandGate analysis of land requirements across energy and technology infrastructure found that utility-scale solar farms require 7 to 10 acres per megawatt, onshore wind projects require 60 to 80 acres per megawatt, and large AI data center campuses require hundreds of acres per 10 megawatts of compute capacity. Aggregated across all sectors, the United States may need between 7 and 13 million acres of new energy- and data-driven land by 2040.
The longer-term response to power scarcity is already beginning to take shape. Modular nuclear reactors are being actively evaluated by several major technology companies as a behind-the-meter power solution that would reduce dependence on grid capacity. Microsoft has announced agreements related to nuclear power development. Amazon has explored similar arrangements. If modular nuclear becomes commercially viable at scale within the next decade, it has the potential to partially decouple data center siting from proximity to existing grid infrastructure — which would alter the geographic distribution of data center land value in ways that are difficult to fully anticipate today.
Energy-as-a-service models — in which power developers guarantee electricity supply to data center operators rather than requiring them to own or manage generation assets — are also emerging as a mechanism for unlocking sites with good land characteristics but inadequate existing grid access.
What This Means for Real Estate Investors in 2026
The data center land rush remains one of the most significant real estate stories in the country. But the environment in which it is playing out has become materially more complex than it was 18 months ago.
The Ratepayer Protection Pledge has shifted the cost burden of infrastructure development more explicitly onto technology companies and developers. State-level regulatory activity is accelerating, with moratorium proposals and revised incentive structures creating meaningful jurisdiction-by-jurisdiction variation in development feasibility. The project pipeline is cooling as grid capacity constraints prove more binding than earlier projections assumed. Electricity prices are rising, affecting both data center operating economics and the political environment in which development approvals are sought.
None of these developments invalidates the fundamental thesis — that power-ready land is scarce, that AI-driven data center demand is structural and durable, and that the premium for sites with existing grid access reflects a real and significant constraint. What they do is raise the analytical bar for investors evaluating specific opportunities.
The sites that will retain and build value through this more complex environment share specific characteristics. They are in jurisdictions with regulatory frameworks that have proven stable and supportive of data center development. They have power infrastructure that is genuinely committed rather than speculatively queued. They have adequate water access. They are held by developers with the capital and operational capacity to navigate the longer and more expensive entitlement processes that the current regulatory environment demands.
The land rush is not over. It is maturing — becoming more selective, more capital-intensive, and more dependent on jurisdictional and regulatory judgment than it was at its most frenetic. For investors who can navigate that complexity, the opportunity remains substantial. For those who cannot, the risks are more visible in 2026 than they were when this story began.
What markets are you watching for data center development activity? And how are you thinking about the regulatory and cost shifts that are reshaping the sector in 2026?
Leave a comment below.