Georgia Power is condemning more than 330 properties in Coweta and Fayette counties to build a 35-mile transmission line for its Project Wansley expansion, with 20 to 30 homes slated for demolition. The $16 billion grid project is one of the largest in Georgia's history, and roughly 80 percent of the new generating capacity will serve data centers, according to Fortune. The utility says condemnation represents less than one percent of its total land acquisitions, but for the families in the path of the 500-kilovolt power line, the disruption is total.
Residents told Fortune that Georgia Power's compensation offers ran roughly $100,000 below their properties' market value. The gap between what the utility is willing to pay and what homes are actually worth has ignited anger and legal challenges, even as the state's energy infrastructure races to keep pace with the insatiable appetite of artificial intelligence.
The expansion is a direct response to the data center boom that has made Georgia one of the nation's largest hubs for cloud computing and AI training facilities. Utilities across the country are scrambling to secure power for thousands of new data centers, many of which consume as much electricity as a mid-sized city. Project Wansley is just one example of a national trend: energy demand forecasts have been revised upward dramatically in the past two years, driven almost entirely by AI and high-performance computing.
The Promise of a Personal AI Agent
Two days before the demolition plans drew national attention, Mark Zuckerberg predicted on Meta's earnings call that within five years, billions of people would each have a personal AI agent handling their finances, health, and relationships. He envisions AI agents that book appointments, manage household budgets, and negotiate on behalf of their users—a layer of intelligence accessible to everyone with a phone. Meta already runs AI agents for more than a million business users on WhatsApp and Messenger.
But the infrastructure required to deliver that promise is already reshaping the lives of people who will likely never use an AI agent. The residents facing condemnation orders are not tech adversaries or anti-progress activists; they are homeowners who have lived on their land for generations. Some have been offered sums that would not even cover the replacement cost of their homes, let alone the lost community ties and disrupted lives that come with forced relocation.
Zuckerberg's pitch is that the pain is temporary and the payoff transformative. Yet the gap between the vision and the present capability is wide. No personal AI agent currently on the market can reliably manage a household budget, let alone the full spectrum of daily life that Zuckerberg described. The industry's track record with ambitious bets is mixed at best.
Meta's Balance Sheet Tells the Story
Meta's own balance sheet reveals the scale of the bet. The company posted revenue of nearly $61 billion in the second quarter of 2026, up 28 percent year over year, but its free cash flow collapsed 91 percent to $784 million as capital expenditure surged. Meta has guided investors toward total spending of $130 billion to $145 billion this year, most of it on AI infrastructure.
That spending spree is staggering even by Big Tech standards. For context, Meta's capital expenditures have more than tripled since 2022, when the company spent roughly $31 billion. The current guidance represents a more than fourfold increase in just four years. The company is betting its entire future on AI, believing that personal agents will become the next massive computing platform, just as smartphones were in the late 2000s.
But investors are skeptical. Meta's stock fell nearly ten percent after the Q2 results, a sign that markets are more moved by collapsing cash flow than by the promise of an agent in every pocket. The company's Reality Labs division has lost roughly $88 billion since 2021 on a metaverse bet that has yet to produce a mass-market product. That history of massive spending without a clear payoff looms over the current AI gamble.
Power Bills Climbing Across the Country
The costs of AI infrastructure are not confined to one utility's service area. In the PJM grid region, which runs from New Jersey to Illinois, capacity charges jumped more than tenfold in a single auction cycle, driven largely by data center demand. Industrial electricity prices rose 31 percent in Pennsylvania and 26 percent in Ohio in one year, against a seven percent national average. These increases are hitting factory towns that have nothing to do with AI, punishing manufacturers and small businesses that cannot easily absorb higher input costs.
Homeowners are feeling the pinch too. Utilities across the country are filing rate cases to recoup the billions of dollars they are spending on new transmission lines, substations, and natural gas plants to serve data centers. Regulators are approving many of these increases, often with little scrutiny, because the data centers represent jobs and tax revenue. But the burden falls disproportionately on ordinary ratepayers, many of whom will never benefit directly from the AI services the infrastructure enables.
Community opposition has grown accordingly. Data centre protests have spread to 42 US states, with more than 140 rallies in a single day of action coordinated by a coalition that cuts across political lines. Conservative groups concerned about land use and eminent domain have joined environmental activists worried about fossil fuel generation. Local officials in Virginia, Ohio, Texas, and Arizona have rejected or delayed data center proposals, and community pressure has blocked or delayed tens of billions of dollars in planned construction.
The Human Cost of AI Progress
The tougher question is whether the human costs of building AI infrastructure can be justified by a product that does not yet exist at scale. For the families in Coweta County facing demolition orders, the promise of a personal AI agent in five years is abstract at best. What is concrete is a 500-kilovolt power line cutting through their properties so that data centers can train the models that might, eventually, produce something useful for ordinary people.
This is not the first time that technological progress has imposed disproportionate costs on local communities. The construction of the interstate highway system in the 1950s and 1960s demolished thousands of homes and displaced entire neighborhoods, often hitting poor and minority communities hardest. Railroads and early electric utilities similarly wielded eminent domain with little regard for property owners. In each case, the benefits of the infrastructure eventually spread to millions, but those benefits came decades later and were never shared with those who lost the most in the interim.
AI infrastructure may follow the same pattern. Data centers are generating massive economic value, and the AI models they train may eventually transform industries from health care to logistics. But the timeline is uncertain, and the immediate costs are real and measurable. The gap between Zuckerberg's vision of billions of personal AI agents and the reality of today's often-anecdotal AI reliability is enormous.
Meta's own track record complicates the forecast. The company has already spent billions on AI data centers and research, yet its most prominent consumer AI products remain limited to chatbot features in WhatsApp and Messenger. The promised agents that can negotiate bills, manage finances, and coordinate relationships are nowhere near production-ready. Meanwhile, the company continues to burn cash at an unprecedented rate.
Fortune reported that Georgia Power's compensation offers were significantly below market value, but utility officials maintain that they are following state law and have offered fair appraisals. The company has noted that condemnation represents less than one percent of its total land acquisitions, painting the forced takings as a rare exception. For the affected homeowners, though, that statistic is cold comfort when the bulldozers are scheduled to arrive.
The broader issue is one of regulatory accountability. States have raced to attract data centers with tax breaks and expedited permitting, often without fully assessing the long-term impacts on electricity rates, water resources, and land use. Virginia and Georgia have become data center meccas and are now grappling with the consequences: soaring power costs, strained grids, and community backlash. Other states may learn from their example, but the incentive to court the next multibillion-dollar facility is powerful.
As the debate continues, the industry has offered few concrete answers. Tech executives talk about next-generation nuclear reactors, small modular reactors, and massive investments in solar and wind, but none of those solutions are ready to power the AI buildout in the near term. In the meantime, utilities are forced to fall back on natural gas and, in some cases, coal to meet demand, undermining climate goals and driving up emissions.
The gap between the vision and the present capability is wide, and it is ordinary families who are paying the upfront price. Whether the trade-off is worth making is a question the industry has not yet been forced to answer. For now, the power line advances, the homes come down, and the bills rise—all in the name of an AI future that remains, for most people, an abstraction.