Eminent Domain, AI Data Centers, and the Erosion of Property Rights

On May 21, 2026, the New York Post published an account of Ansley Brown, a 27-year-old woman from Coweta County, Georgia, who is fighting to keep her mother from being forced to sell the home where Brown grew up. The buyer is Georgia Power. The mechanism is eminent domain. The reason: a 35-mile, 500-kilovolt transmission corridor that Georgia Power calls Project Wansley, which will deliver electricity to at least four AI data centers currently under development in Coweta and Fayette counties. At least 330 properties are in the path. Between 20 and 30 homes are scheduled for outright demolition. Hundreds of other landowners are being asked to grant permanent easements that will plant 500-kilovolt power towers feet from their bedroom windows.

The Georgia Power case is the most visible single example of a pattern that the framework's housing analysis (Articles 7, 17, 18, and 19 of this catalog) has been identifying as a specific saleability risk: the structural erosion of American property rights through the expanded use of eminent domain in service of private infrastructure development, accelerated in the current period by the electrical demands of artificial intelligence compute. The pattern has been building for two decades. The framework's reading is that it has now reached an inflection point where the specific cases are visible enough, the legal structure is settled enough, and the underlying infrastructure pressure is sustained enough that the pattern itself becomes a primary factor in housing saleability across affected geographies.

This essay is the sixth installment of the Watching the Cracks series. It does three things. First, it traces eminent domain from its origins in the Magna Carta through the U.S. Constitution to the Supreme Court's 2005 Kelo decision and the subsequent state-level responses. Second, it examines the current Georgia Power case and the broader pattern of AI-infrastructure-driven eminent domain expansion across multiple states. Third, it examines the New Albany Company's privatized-governance model in central Ohio as the most developed example of what happens when the eminent-domain machinery is integrated into a private real estate development apparatus rather than operated through traditional municipal channels. The framework's reading runs through all three: housing's saleability is structurally contingent on the state's restraint in weaponizing property rights, and the cases now visible in the data show that restraint eroding in real time.

The legal history

The fundamental claim that a sovereign authority can compel the sale of private property for public purposes is older than any Anglo-American legal tradition. It appears in the Old Testament (the prophet Ahab's seizure of Naboth's vineyard is condemned, suggesting the prohibition was operating against the practice). Roman law recognized the principle. By the medieval period, European sovereigns routinely exercised the right to take land for fortifications, roads, and public works.

The first legal restriction on the sovereign's eminent domain authority in the Anglo-American tradition appears in the Magna Carta of 1215. Article 28 of the original charter required that "No constable or other royal official shall take corn or other movable goods from any man without immediate payment, unless the seller voluntarily offers postponement of this." Article 39 provided that "No freeman shall be seized or imprisoned, or stripped of his rights or possessions... except by the lawful judgment of his equals or by the law of the land." The Magna Carta did not abolish the king's ability to take property — it conditioned that ability on payment and due process. The framework's reading: the principle that private property cannot be taken arbitrarily, but can be taken with compensation under defined procedures, is the operational structure that has governed the question ever since.

The Magna Carta's restrictions evolved through English common law over the following five centuries. By the time of the American founding, the principle that government could take private property for public use but must pay just compensation was well established in legal theory but not consistently in practice. The Vermont Constitution of 1777 and the Massachusetts Constitution of 1780 were the first written documents to formally require compensation for eminent domain takings — predating the U.S. Constitution by a decade.

The federal Takings Clause appears in the Fifth Amendment to the U.S. Constitution, ratified in 1791. The clause reads, in its entirety: "nor shall private property be taken for public use, without just compensation." Two specific limitations are encoded. First, the taking must be for public use — not for the benefit of any particular private party. Second, the owner must receive just compensation — typically interpreted as fair market value. The Fifth Amendment was inserted by James Madison and was not the product of public campaign or popular demand; eminent domain protections had not been a major theme in the Anti-Federalist critique of the unamended Constitution. Madison included it because he believed property protection was foundational to constitutional governance, not because the public was loudly asking for it.

The "public use" limitation governed eminent domain doctrine for approximately 150 years. Roads, schools, military installations, post offices, railroads, public utilities operating under common-carrier obligations — these were the canonical "public use" cases. Where the government took property and operated the resulting infrastructure itself, or where it transferred to a private party operating under public-utility obligations, the taking was upheld. Where the taking transferred property from one private party to another for purely private benefit, courts generally declined to uphold.

This structure began to weaken in the post-WWII period. Berman v. Parker (1954) upheld an urban renewal taking in Washington, D.C. where the property was being transferred to private developers as part of a broader slum clearance plan. The court accepted the argument that the broader plan served public purposes (eliminating blight) even though specific properties were being transferred between private hands. Hawaii Housing Authority v. Midkiff (1984) upheld a Hawaii statute that broke up large land oligopolies by transferring land from large landowners to their tenants — again, private-to-private, justified by the broader public purpose of reducing land concentration.

The decisive doctrinal break came in Kelo v. City of New London (2005). The case arose in New London, Connecticut, where the city used eminent domain to seize Susette Kelo's pink waterfront house and several neighboring properties. The city's plan was to transfer the land to a private developer who would build commercial and residential developments adjacent to a new Pfizer research facility, with the stated public purpose being economic development and tax revenue generation. Kelo and her neighbors sued, arguing that taking property to transfer it between private parties for economic development purposes did not satisfy the Fifth Amendment's "public use" requirement.

The Supreme Court ruled 5-4 against Kelo. Justice John Paul Stevens, writing for the majority, held that economic development qualified as "public use" under the Takings Clause — that the general benefits the community would enjoy from economic growth were sufficient. The "public use" requirement, in the majority's reading, did not require that the property be used directly by the public; it required only that the taking serve a "public purpose," with that purpose interpreted broadly. Justice Sandra Day O'Connor's dissent, joined by Chief Justice Rehnquist and Justices Scalia and Thomas, argued that the majority's reading "effectively erases" the public use requirement and that under the new doctrine, "the specter of condemnation hangs over all property... Nothing is to prevent the State from replacing any Motel 6 with a Ritz-Carlton, any home with a shopping mall, or any farm with a factory." Justice Thomas wrote a separate dissent arguing that the majority's reading was textually indefensible.

The Kelo decision sparked a political backlash that has no real parallel in modern eminent domain history. Forty-seven states subsequently passed legislation or constitutional amendments restricting eminent domain takings for economic development purposes. Twelve states amended their state constitutions specifically. Eleven state supreme courts either strengthened their property rights protections or rejected the federal Kelo analysis as a matter of state constitutional law. The Institute for Justice, which represented Kelo in the Supreme Court case, has subsequently helped save approximately 20,000 homes and small businesses from condemnation through community organizing and litigation.

The Kelo property itself was never developed. The private developer was unable to obtain financing and abandoned the project. As of 2026, the land where Susette Kelo's pink house once stood is an undeveloped empty lot.

The current pattern

What the framework reads in the current eminent domain expansion is that the state-level Kelo backlash, while substantial, did not address a category of takings that has emerged with full force only in the past three years: takings for utility infrastructure that primarily serves private industrial customers, with the public-use justification supplied by the utility's regulated status rather than by the direct public benefit of the infrastructure being constructed.

Georgia Power's Project Wansley is the canonical case. The project is a 35-mile, 500-kilovolt transmission corridor. The legal authority for Georgia Power to invoke eminent domain comes from its status as a regulated utility under Georgia state law — the same statutory framework that would authorize Georgia Power to take property for transmission lines serving residential neighborhoods, hospitals, or schools. The actual end-use of the electricity that will flow through the Project Wansley lines is, by Georgia Power's own statements, "at least four AI-driven data centers" plus general grid reliability improvements. The Georgia Public Service Commission has approved plans for approximately 1,000 miles of additional transmission line construction over the next decade, driven primarily by data center demand. Metro Atlanta is now the fastest-expanding data center market in the United States and ranks among the top three globally.

The framework's reading is that this is Kelo's structural mechanism extended one step further. In Kelo, the government took property for transfer to a private developer who would build private commercial buildings, with the public-use justification provided by general economic development. In Georgia Power's Project Wansley, a regulated utility takes property for transmission lines that will primarily serve specific private corporate customers (the data center operators), with the public-use justification provided by the utility's regulated status and the general grid-reliability benefits. The legal structure is different. The substantive effect on the affected property owners is identical — their land is taken, against their will, for infrastructure that primarily benefits private parties.

The Ansley Brown case demonstrates the mechanism precisely. Brown's mother purchased the house in Coweta County in 2003, when Brown was five years old. Brown grew up in the house. The house is now in the path of a 500-kilovolt transmission line that will deliver electricity to data centers operated by companies whose existence Brown's mother could not have anticipated when she purchased the property. Georgia Power has offered Brown's mother 125% of fair market value for the property — but the proceeds are in the form of a 1099-S check that creates immediate tax liability if not rolled into another property purchase within a defined period, and the family is being asked to vacate by a specific date determined by the utility's construction schedule rather than by the family's own planning. The framework reads this as a forced transaction with terms set by the taking party — the textbook definition of an eminent domain action, with the additional feature that the underlying public-use justification has effectively dissolved into "AI compute demand requires more transmission capacity."

The pattern is not limited to Georgia. Project Sail, an 829-acre hyperscale data center campus proposed by Prologis/Atlas in Coweta County, was approved by a 3-2 county commission vote that rezoned conservation land to industrial use, with local residents subsequently suing. Similar disputes are visible in northern Virginia (Loudoun and Prince William counties, where the original "Data Center Alley" cluster has expanded into rural areas with active eminent domain proceedings), in Texas (where the ERCOT grid is racing to add transmission for data center loads), and in Arizona (Maricopa County and beyond, where the same dynamic is playing out at smaller scale). The framework's reading is that the AI compute boom is producing eminent domain pressure on rural and suburban properties at a scale not seen since the interstate highway construction era of the 1950s-1960s, with the legal structure governing the takings substantially looser than what governed those earlier infrastructure projects.

Georgia Power's response to the criticism has emphasized that eminent domain "is a last resort" and "comprises less than 1 percent of all of the land transactions each year" and that the company offers compensation "well above market value." The framework's reading: each of these statements is technically true and analytically incomplete. Eminent domain is a last resort in the formal procedural sense; the company first negotiates voluntarily, and the formal condemnation process is invoked only when negotiations fail. But the threat of condemnation is present from the first contact with the landowner, and the formal procedural protection that eminent domain offers (court oversight, just compensation, due process) provides limited substantive protection when the landowner does not want to sell at any price.

The New Albany model

The Georgia Power case is the most visible current example of eminent domain in service of AI infrastructure. The most developed example of the broader pattern — where private development capital is integrated with municipal authority into a unified governance apparatus — is the New Albany Company's operation in central Ohio. The framework's reading is that the New Albany model is what the Georgia Power pattern looks like after thirty years of compound development, and it deserves careful examination because of what it reveals about where the broader trajectory is heading.

The history begins in 1986. Leslie Wexner, the founder of L Brands (parent of The Limited, Express, Victoria's Secret, Bath & Body Works, and Abercrombie & Fitch), and Jack Kessler, an Ohio business and civic leader, were reportedly driving through central Ohio in Wexner's Land Rover when they identified large tracts of undeveloped farmland near the village of New Albany — at the time a community of fewer than 500 people on the northeastern edge of Franklin County. Wexner is reported (per the Cleveland Plain Dealer) to have described the area as his "future kingdom." He and Kessler formed the New Albany Company that year as a private real estate development entity.

The acquisition strategy that followed has been documented in multiple investigative pieces over the years and most recently in coverage by Tech Policy Press, The American Prospect, Unlimited Hangout, and The Ohio Register. The New Albany Company assembled large tracts of land through purchases from local farmers, often at prices above market but subject to non-disclosure agreements that prevented the sellers from discussing the transactions publicly. Paper corporations were established to obscure the ownership chain. The Cleveland Plain Dealer reported that the corporations were "spun off to obscure ownership." A strict master plan was imposed on the assembled land — Georgian architecture, white fences, country-club aesthetics, residential density limits — that the company has continued to enforce on subsequent development across the area.

Over the following thirty years, the New Albany Company assembled the 12,000-acre New Albany International Business Park, of which approximately 9,000 acres sit within the City of New Albany's boundaries. The village's population grew from fewer than 500 to approximately 11,000. The business park became one of the largest concentrated commercial developments in the Midwest. Abercrombie & Fitch, originally headquartered in the park, was for many years responsible for approximately 75% of the city's income tax revenue.

The data center boom transformed the park. As of 2026, the business park hosts more than 40 data centers operated by Amazon Web Services, Google, Meta, Microsoft, and Intel, with additional facilities under construction. The Intel semiconductor facility — announced in 2022 at $20 billion and subsequently expanded in scope to approximately $28 billion — sits on 3,190 acres of land that was annexed to New Albany from Jersey Township, just across the county line in Licking County. The annexation process involved the New Albany Company approaching landowners in and around Johnstown, Ohio, offering well above market rate (in many cases over $1 million per property), subject to non-disclosure agreements. The transactions were completed before the Intel deal was publicly announced. JobsOhio, the state's quasi-private economic development entity, secured a $150 million grant that paid for the land Intel ultimately built on, as part of a state incentive package totaling more than $2 billion.

J.P. Nauseef, JobsOhio's CEO, told Forbes of the Intel deal: "It wouldn't have happened without them" — referring to the New Albany Company. The framework reads this as straightforward fact rather than indictment. The Intel facility, the data center expansion, the broader transformation of central Ohio into the "Silicon Heartland" — these required a private entity capable of assembling the land at speed, with the operational capacity to coordinate municipal annexation, state incentive packages, and federal infrastructure funding. JobsOhio is a quasi-private entity; the New Albany Company is a private entity; the Columbus Partnership (founded by Wexner and Kessler in 2002, with roughly 70 corporate members) functions as the de facto regional economic development authority across 11 central Ohio counties. The Wexner-Kessler apparatus is the operational governance layer for the region's industrial development.

The framework's analytical observation is direct. In central Ohio, the entities that perform the functions traditionally associated with municipal government — land use planning, infrastructure coordination, economic development incentive negotiation — are private corporations whose decision-making is not subject to the public accountability mechanisms that govern actual municipal government. This is not a Kelo-style case where a municipal authority transfers property to a private developer; it is the further structural development where the planning and acquisition functions have themselves been migrated from public to private institutions, with municipal authority operating in a coordinated but subordinate role.

A 2025 Forbes investigation reported that Wexner personally made approximately $2 billion from investments in CoreWeave, the AI infrastructure and cloud computing firm that is among the major customers driving central Ohio data center demand. The framework does not need to make any claim about the propriety of this investment — Wexner is a private individual entitled to invest as he chooses. The framework's structural observation is that the same individual who controls the private entity that performs municipal-governance functions for the region is personally a major beneficiary of the AI infrastructure investment that the region's governance has been organized to facilitate. The alignment is operational, not necessarily conspiratorial; the operational alignment is what the framework's analysis is required to engage.

The Wexner-Epstein connection — Jeffrey Epstein served as Wexner's "personal money manager" for years, with documented financial and real estate ties — appears in essentially every credible piece of journalism about the New Albany model. The framework's reading is that this connection is relevant context but not analytically central to the structural pattern this essay is examining. The pattern of privatized municipal governance and integrated eminent-domain-adjacent land assembly would be structurally identical, in the framework's reading, regardless of the specific biographical features of the entities involved. The reader who wants to engage the Wexner-Epstein material can do so through the existing journalism; the framework's analysis here is about the governance pattern itself.

What the framework reads in the pattern

The framework's housing analysis from Article 7 of this catalog identified six factors that determine the saleability of housing as an asset class: divisibility, durability, transportability, homogeneity, widespread demand, and freedom from political weaponization. The first five are intrinsic to the asset itself. The sixth — freedom from political weaponization — is contingent on state restraint in invoking the various mechanisms by which property rights can be impaired or transferred against the owner's will.

The Georgia Power case demonstrates that this sixth factor is now operating against affected property owners at scale. Ansley Brown's mother does not want to sell her home. The legal mechanism by which Georgia Power can compel the sale exists because the state legislature has delegated condemnation authority to the regulated utility. The public-use justification has dissolved into the indirect chain of reasoning that AI compute demand requires more transmission capacity which requires more land acquisitions which are pursued through condemnation when voluntary negotiation fails. The chain is legally adequate. It is not structurally identical to the public-use justification that originally constrained the takings authority.

The New Albany case demonstrates the next structural step. The acquisitions there have not generally proceeded through formal eminent domain — the New Albany Company has used premium-price offers with non-disclosure agreements to assemble land without invoking the legal coercion machinery. The framework's reading is that this is not an argument for the absence of coercion. The structural reality is that landowners facing a private development entity with effectively unlimited capital, operating in coordination with municipal authority that can deny rezoning, with state-level political support that can structure incentives against non-cooperation, face economic pressure that operates substantively like eminent domain even when it does not require the formal condemnation process. The non-disclosure agreements prevent collective action by sellers; the master plan removes substantive options for future use; the alternative to selling at the offered price is to remain in a neighborhood being systematically redeveloped around the holdout. The legal structure is technically voluntary; the substantive position of the landowner is closer to compelled.

Five framework observations follow.

First, the saleability of American housing is now meaningfully impaired by weaponization risk in geographies where AI infrastructure expansion is concentrated. The framework's Article 7 prediction that this factor would emerge as a binding constraint was made in the abstract; the Georgia Power case demonstrates it operating empirically. Where transmission infrastructure expansion is pending, housing prices in the affected corridors face downward pressure beyond what fundamental supply-and-demand dynamics would predict. The framework's metro saleability map from Article 17 will need to be updated in subsequent installments to reflect this factor.

Second, the post-Kelo state-level reforms are inadequate to address the current pattern. The forty-seven states that strengthened eminent domain protections after 2005 generally focused on the specific Kelo issue — economic development takings that transfer property between private parties. The current pattern operates through delegated utility condemnation authority, which the post-Kelo reforms did not substantially address. The framework's prediction: state-level legislative responses will emerge over the next 2-4 years targeting the utility-delegation mechanism specifically, but the current property owners facing condemnation will not benefit from those reforms in time.

Third, the New Albany model is structurally exportable. The integrated apparatus of private land assembly, municipal annexation, state-level incentive coordination, and political alignment that the New Albany Company assembled over thirty years is not unique to central Ohio. Comparable structures are emerging in northern Virginia, the Phoenix metro, the Dallas-Fort Worth metroplex, and other regions where AI infrastructure investment is concentrated. The framework predicts that the next decade will see at least 3-5 additional regional examples of the New Albany pattern reaching similar structural maturity, with similar governance implications.

Fourth, the eminent domain pattern interacts with the housing carrying-cost wedge identified in Article 19 to compound saleability impairment in specific metros. A property owner facing a 500-kilovolt transmission line easement, in a metro where property tax assessments have surged 45-65% since 2019 and where insurance premiums are escalating beyond inflation, is exposed to three independent saleability impairments simultaneously. The framework's reading is that these factors are not independently additive; they are interactive in ways that compound the saleability impact beyond the sum of the individual factors.

Fifth, the framework's broader claim about substitute layers from earlier in this catalog now extends to the institutional substrate of property rights themselves. The same way agency MBS (Article 8) operates as a paper substitute for the underlying low-saleability properties of residential mortgages, the regulated utility's condemnation authority operates as an institutional substitute for the original constitutional structure that constrained eminent domain to public uses. The substitute structure has held, in framework terms — the takings are formally legal, the procedural protections operate as designed, the compensation requirements are met. The substantive constraint that the Fifth Amendment was originally designed to provide has been substantially weakened in operation.

What households should take from this

The framework's specific operational observations for household readers making housing decisions in geographies affected by AI infrastructure expansion:

Diligence on transmission corridor exposure should now be standard. Any household purchasing property in Georgia, northern Virginia, central Ohio, Phoenix metro, Dallas-Fort Worth, or other AI infrastructure expansion areas should review the relevant state public service commission filings for proposed transmission line corridors. Property within several miles of an approved or proposed corridor faces meaningful saleability exposure that does not appear in the standard real estate disclosure documents.

The post-Kelo state reforms should be reviewed for their specific scope. Some states (Florida, Mississippi, Oklahoma) have meaningfully restricted economic development takings but left utility takings largely untouched. Others (New Hampshire, Michigan) have been more restrictive across the board. The relevant question for a specific property is whether the state's reforms cover the kind of taking the property might face — typically a utility taking — not whether the state has eminent domain reform on the books generally.

The local political economy of land use planning matters more than it traditionally did. Where municipal authority operates in close coordination with a private land assembly entity (the New Albany model), the household's protection against unwanted development depends on the underlying political relationships rather than on formal regulatory protection. The framework's recommendation: households investigating long-term property holdings should examine the relationships between local government and major private development entities as part of standard due diligence.

Settled doctrine is not the same as stable doctrine. Kelo has been "settled" since 2005 in the sense that the Supreme Court has not revisited the underlying analysis. It is not stable in the sense that the constitutional theory it endorsed continues to be widely contested, multiple state supreme courts have rejected it as a matter of state constitutional law, and the political coalitions that produced the post-Kelo backlash remain active. The framework predicts that Kelo will be revisited at the Supreme Court level within the next decade, with substantial uncertainty about whether the revisiting will reinforce or weaken the current framework.

The closing observation

In 1215, the barons forced King John to sign a document at Runnymede that required the king to compensate his subjects for taken property and to operate within defined legal procedures when invoking that authority. The principle survived through eight centuries of English and American legal development, was encoded in the Fifth Amendment of the U.S. Constitution, and operated as the foundational structure of American property rights through most of American history.

In 2005, in a 5-4 decision, the Supreme Court substantially weakened the "public use" limitation that had constrained the takings power. In 2026, the structural consequences of that weakening are visible in concrete cases — Ansley Brown in Coweta County, the Maszk family in Fayette County, the hundreds of other Georgia Power transmission corridor cases, the broader pattern across Virginia, Texas, Arizona, and central Ohio. The framework's analytical claim is not that the current pattern is unprecedented in American history — the railroad era and the interstate highway era both involved comparable property rights tension. The claim is that the combination of post-Kelo doctrinal weakening, AI infrastructure scale, integrated private-public development apparatus (New Albany model), and the broader monetary architecture's pressure on housing as the primary household wealth asset, has produced a structural environment in which property rights are less secure in operation than they have been at any point since the Magna Carta encoded the original constraint.

This is the framework's reading. The cases are visible. The pattern is empirical. The constitutional structure that was supposed to prevent this has substantially failed in operation. The framework's job is to make the failure visible while there is still time for a constructive political response to address it.

The next installment of Watching the Cracks will engage the closures of small private colleges across the Northeast and Midwest — Hampshire College's closure announcement on April 14, 2026, the broader pattern of demographic and financial pressure on small institutions, and the framework's reading of what AI specifically is doing to the value proposition of the traditional four-year degree. The watching continues. The geography of what is being watched now includes the rural and suburban corridors where AI infrastructure is being built — and the property owners along those corridors whose saleability has been impaired by structural forces they had no role in producing.

---

This is the sixth installment of "Watching the Cracks." The framework's predictions recorded here for future testing: at least 3-5 additional regional examples of the New Albany privatized-governance model will reach structural maturity over the next decade; state-level legislative responses targeting utility-delegation eminent domain will emerge within 2-4 years but will not benefit current property owners facing condemnation; the Kelo decision will be revisited at the Supreme Court level within the next decade. The Wexner-Epstein historical connection is documented in multiple credible sources but is not analytically central to the governance-pattern analysis this essay engages.