A small Nebraska village has introduced an uncomfortable question into the economics of AI infrastructure: whether a site can retain economic value by never becoming an industrial site at all. Nehawka amended its zoning code in August 2026 to prohibit data centers across its zoning districts, including associated substations, transformers and generators. The decision matters beyond the physical boundaries of a community because infrastructure development often treats developable land as an option that can be exercised when technical and economic conditions align. A developer sees that option through power availability, connectivity, expansion potential and the ability to assemble a large enough site for future demand.

An infrastructure investor can therefore assign value to what a site could become even before construction begins. Nehawka introduces the opposite proposition: some economic value may come from preserving what a site will not become. That proposition challenges a common assumption behind AI infrastructure development without requiring an argument against AI, computing or industrial investment. The question becomes whether economic optimization should always mean converting the maximum amount of land into the highest-intensity use available.

When Development Stops Being the Default

The most provocative implication is that development does not have to represent the only economically rational end state for an attractive site. In a rapidly expanding AI infrastructure market, undeveloped land can carry strategic value when future demand supports larger electrical loads, larger buildings and additional capacity. That logic can create an implicit hierarchy in which a site’s development value rises with its potential industrial intensity. Yet that calculation can leave out a different consideration: permanence. A community that decides a site should remain agricultural, residential, open or otherwise low-intensity can be understood as placing value on the absence of future conversion.

That value may not appear neatly in a conventional land-acquisition model because those models typically focus on the economic potential of acquiring and developing a site. Nehawka therefore offers a useful case study for considering whether infrastructure economics should account for situations in which the strongest site decision involves eliminating a future development pathway. The decision does not require declaring AI infrastructure undesirable. It requires recognizing that a community can regard continued non-conversion as a deliberate land-use outcome rather than simply a failure to capture investment.

The Site Option Has a Cost

Site control can provide flexibility, but that flexibility can carry a hidden economic assumption: that the owner and surrounding jurisdiction will continue to permit the intended conversion. That assumption becomes less reliable when communities begin assigning value to keeping land in its existing form. A site can possess transmission access, transportation proximity and physical scale while still carrying an economic identity that does not align with industrial expansion. Once that identity becomes durable, the theoretical upside of conversion can become less relevant where existing land use carries greater long-term value than redevelopment. This creates a new question for infrastructure planners: how much of a site’s theoretical development value actually belongs in an investment thesis if conversion will never occur?

The answer matters because AI infrastructure projects increasingly compete for large, strategically positioned sites rather than simply seeking any available acreage. A site that cannot convert may appear less attractive under a conventional development model while remaining economically rational when its existing long-term land use is valued. The resulting tension is not necessarily between technology and land but between two competing definitions of what makes land valuable.

Choosing What Never Happens

The most interesting part of the Nehawka case is not the ban itself but the economic question raised by making a long-term choice about what a place will not become. Infrastructure investment often values optionality because optionality allows capital to respond to future demand. Yet an emphasis on optionality can also encourage investors to view strategically positioned sites primarily through their future development potential. Nehawka presents a counterexample in which removing a development option can itself become the intended land-use outcome. That logic deserves attention because AI infrastructure increasingly depends on physical assets whose useful lives extend far beyond individual technology cycles.

If a community determines that retaining its existing land-use identity has lasting economic value, then the absence of a data center can become part of how the site’s long-term value is understood. The industry may eventually need to consider that factor alongside electrical capacity and expansion potential when evaluating a site’s long-term viability. As AI infrastructure expands, it becomes increasingly important to recognize that not every site will be available or appropriate for conversion.

The Next Site Question Is Not Always Where to Build

A conventional infrastructure question asks where the next data center should go, but Nehawka suggests a harder question: where should development deliberately stop? That question does not reject growth or diminish the importance of AI computing. Instead, it recognizes that infrastructure economics operate inside a finite physical landscape where different forms of value can coexist. A site can be valuable because it supports computing, because it supports another established use or because its long-term value lies in remaining outside industrial conversion and speculative redevelopment.

Those outcomes can carry different economic consequences across decades. The challenge for infrastructure leaders is to avoid assuming that the highest-intensity use automatically represents the highest-value use. Nehawka’s prohibition makes that assumption unusually visible because it removes data-center development as an available land-use pathway within the village’s zoning jurisdiction. The larger lesson is that AI infrastructure may eventually need to account for the value of land that remains in its existing use.

The Strongest Infrastructure Decision May Be a Refusal to Convert

Nehawka’s decision ultimately points toward a different way of thinking about infrastructure economics, one that asks whether permanence can carry value rather than treating it only as an obstacle. The community has created a clear boundary around a future it does not intend to pursue, and that boundary changes the development possibilities for sites inside its zoning jurisdiction. For the AI infrastructure industry, the important question is not whether similar decisions will become widespread but whether the industry may need to account for places that deliberately remove data-center conversion from their available land-use options. That possibility makes non-conversion a potentially relevant consideration in site strategy, investment underwriting and long-term infrastructure planning.

It also gives end users a reason to consider an issue that can otherwise appear distant from computing economics. Reliable AI capacity depends not only on finding sites that can support massive infrastructure but on avoiding sites whose long-term assumptions cannot hold. The industry’s next generation of site strategy may therefore need to consider a category that receives little attention today: land whose long-term value may come from remaining outside AI infrastructure development. In that context, the most consequential site decision may not always be choosing where to build, but accurately recognizing where building should never enter the model.