Every time you visit a news website, you are likely to encounter claims about how artificial intelligence (AI) is transforming the economy. However, an important aspect often gets overlooked: Who bears the financial burden of maintaining the energy-intensive infrastructure that supports these advancements?
Beneath the surface of user-friendly chatbots and intrusive workplace monitoring systems lies a network of industrial-grade data centers that operate tirelessly. These data processing hubs facilitate endless queries, analyze vast amounts of data, and engage in continual AI training cycles—enabling late-night inquiries to ChatGPT about procrastination remedies.
Constructing a single hyperscale data center can cost approximately $11 billion. The funding typically comes from tech giants like Meta Platforms, Microsoft, Google, and Amazon, or from private ventures securing contracts with these companies. In Virginia alone, there are around 300 hyperscale data centers, each contributing significant costs related to land acquisition, energy consumption, and the expansion of infrastructure.
Yet here’s the catch: While it may seem that Silicon Valley is shouldering these expenses, the financial repercussions extend far beyond their domain. First, you might notice a gradual increase in your electricity bill. Then, the consequences ripple through your daily life, resulting in higher rents and elevated borrowing costs. Before you know it, a general rise in prices occurs without any clear explanation.
The cost of living has indeed increased, despite promises of technological prosperity.
The Price of Power Hunger
The scale of data centers required for AI is staggering. As physicist Joseph Romm from the University of Pennsylvania’s Center for Science states, “These are not just typical facilities; they are vast, energy-intensive complexes where a single hyperscaler can have an energy footprint akin to that of an entire country.”
When energy consumption escalates to such levels, it impacts the broader economy.
Large data centers often negotiate advantageous deals with utility companies, allowing them to draw cheaper power in exchange for reducing consumption during peak times. This arrangement sounds practical on the surface, but it overlooks a crucial fact: the power grid must be equipped to handle their maximum consumption at all times. This means the need for additional power plants, extra infrastructure, and spreading these costs across consumers’ utility bills.
For example, a significant report on the PJM power grid, which services much of the eastern United States, revealed that regular utility customers in seven states effectively covered about $4.3–$4.4 billion in transmission upgrades. A considerable portion of this expense was driven by data centers, but rather than billing the tech companies directly, the costs were absorbed by every customer’s electric rates.
Virginia residents are acutely aware of these dynamics. A Bloomberg analysis showed that by 2024, data centers already consumed nearly 40% of the state’s electricity, primarily due to the numerous mega-facilities in Northern Virginia. A state-commissioned review projected that the swift growth of data centers might lead to an increase of about $14 to $37 monthly for the average household electric bill in the coming years. They expand; you pay.
While Virginia leads the nation in data center establishments, Texas is rapidly catching up, with Georgia, Arizona, and Ohio emerging as significant contenders for new AI infrastructure. Residents in these areas are expressing concern as electricity prices rise at double the inflation rate.
Small businesses find themselves paying more for the same operations, even though the demand causing these costs originates from data centers that may be invisible to them. They utilize the same energy, but their bills continue to increase.
Farmers are also feeling the strain as pollution and water demands impact soil quality and groundwater, continually raising input expenses, making irrigation costlier, and shifting more environmental burdens directly onto their farms.
In essence, no one is spared. A recent study by economist Servaas Storm indicates that rising energy costs contribute to inflation throughout the economy. When inflation accelerates, central banks intervene, keeping interest rates elevated for prolonged periods to stabilize the situation. These higher rates become apparent to everyday consumers in the form of increased expenses for mortgages, car loans, and credit cards, thus tightening household budgets.
Increasing housing demand can drive rents and property values up near data centers. Funds allocated for grid enhancements and tax incentives related to these facilities detract from essential community resources—such as schools, public transit, local infrastructure, and other services that help make life affordable and stable.
Even if you’ve never ventured into the realm of AI, you will inevitably contribute to its expenses.
Trickle-up Economics
Wealth generated from AI tends to flow in a familiar direction, with large corporations and their executives reaping the majority of the rewards.
Energy companies are discovering that AI presents a goldmine for the power industry, as windfall-style profits flow to regulated utilities that are making billions from new infrastructure expansions to meet soaring electricity demands. In tandem, the compensation for top executives is also surging.
Some industry groups and analysts attribute rising prices in the data center era to “NIMBYism,” or local opposition to development projects. They argue that the AI-driven demand is outpacing supply due to community pushback, resulting in delayed grid expansion and surging costs. Conversely, policy groups like Third Way assert that a slow approval process is the fundamental issue for consumers, asserting that community resistance leads to higher costs.
Thomas Ferguson, Director of Research at the Institute for New Economic Thinking, finds this argument lacking. “The notion of NIMBYism is vastly exaggerated,” he pointed out. “It’s reminiscent of the Reagan-era ‘snail darter’ controversy, where a minor endangered species became the scapegoat in public debates while the true issues were far more complex.”
He argues that these narratives miss the point that infrastructure battles are often about financial power and competition rather than genuine local dissent, citing past instances where railroads used environmental concerns to thwart improvements that would have increased competition from barges.
“Affluent homeowners in locations like Martha’s Vineyard or California sometimes oppose projects to preserve property values,” Ferguson conceded, “but those situations are uncommon.” He believes that the significant pressures inflating utility prices arise “mostly from the utilities themselves, often with fossil fuel and renewable energy producers colluding against outside competition.”
In summary, energy outcomes are shaped less by “the market” and more by organized business interests that sway policy and investment behind the scenes. As Servaas Storm points out, the lofty promises companies make about AI—its potential to usher in a new era of abundance or to resolve pressing human issues—remain exceedingly difficult (if not impossible) to substantiate.
Nevertheless, tech giants have honed the skill of securing expedited approvals for new data centers, often with state and local officials eager to pursue the prospect of “investment.” However, the true financial reality presents a stark contrast. While there may be some temporary construction jobs and local hiring boosts, research consistently indicates that the long-term economic benefits for surrounding communities are limited, with most advantages accruing to the corporations themselves while localities bear the infrastructure and utility expenses.
Mark Glick, a law and economics professor at the University of Utah, expressed concern over the imbalance: “Companies ought to cover all the additional resources they consume—why should we shoulder the burden?” He insists, “If you’re contributing to peak load, you need to compensate for the difference.”
Joseph Romm raised a red flag regarding the financing of data centers: “Some companies, including Meta, are creating separate entities to own the infrastructure, effectively putting data centers off their balance sheets.” This means the tech firms are not directly accountable for these substantial, long-term debts, facilitating easier investment decisions. However, it also implies that if the economic conditions for AI shift or demand wanes, companies could potentially walk away from substantial financial liabilities, a situation that could result in “tens of billions of dollars” in losses, according to Romm.
Additionally, the source of energy impacts its affordability dramatically. Regions reliant on fossil fuels for energy can find demand from data centers enhancing dependency on systems prone to price volatility. These dynamics can create instability in electricity pricing for households and small businesses. Romm notes that economic and environmental concerns are intertwined, with communities experiencing the local impacts of energy production while costs fluctuate simultaneously.
“Five years ago, major tech corporations were pledging to achieve net-zero emissions and positioning themselves as climate leaders,” he reflected. “Now, they are constructing infrastructure that significantly amplifies electricity demand, much of which is still met by fossil fuels.” He acknowledged that while Microsoft, Google, Amazon, and Meta appear varied in sincerity and execution, they are all adding pressure to the grid demand that contradicts their earlier pledges.
The Politics of Power: Just Say No?
In the 2019 documentary iHuman, Ilya Sutskever, co-founder and former Chief Scientist of OpenAI, remarked, “I think it’s pretty likely the entire surface of the Earth will be covered with solar panels and data centers.”
Given the immense scale of infrastructure envisioned by AI researchers, proactive measures are essential.
Policy must stay ahead of this expansion, ensuring that decisions related to energy, land use, job creation, environmental effects, and AI’s purpose prioritize the well-being of everyday people rather than merely catering to the convenience of powerful corporations and their leaders.
Darren Bush from the University of Houston cautioned in an email that “companies have a history of misleading local governments regarding the benefits of development, and data centers are no exception.” He emphasized that we should avoid a scenario where the expenses associated with data centers are transferred onto citizens. He noted the additional costs involved, which encompass not only electric bills but also environmental impact, water usage, and public health.
It becomes challenging to maintain health—or employment—when sleep is disrupted by the noise of cooling systems and generators, as residents around Virginia’s “Data Center Alley” are currently experiencing.
For Darren Bush, the concerns surrounding data centers are significant enough that a default approach should be one of caution. He believes cities should sometimes simply “just say no” to these projects, as they “bear all the costs with none of the benefits.” If they do proceed, he argues that developers should be held fully accountable. “If they intend to engage with potentially harmful projects, they should ensure that all data center-related externalities are properly taxed,” which would compel those advocating for data centers to reconsider.
Bush stresses that this issue extends beyond city jurisdictions. “State regulators must also be involved,” he adds, underlining that “utility rates for established residents should not be adversely affected by spikes caused by data center demands.” He candidly acknowledges the challenges of actualizing these goals: “That’s all easier said than done, of course,” he concedes. “Which is why I advocate for outright rejections.”
Different policy approaches can steer outcomes in vastly different directions. For instance, mandating that data centers utilize renewable energy sources can influence both long-term pricing and environmental effects. Thoughtful investments in the grid can maintain power supply while preventing your bills from escalating too quickly. Enhancing transparency around incentive agreements and infrastructure costs can provide a clearer understanding of the prevailing situation. With community involvement at the forefront, local residents can help decide which developments move forward and prioritize benefits for the community.
Mechanisms like community benefit agreements empower residents to link large-scale infrastructure initiatives to local needs. These arrangements can encompass job training, infrastructure improvements, or measures to stabilize costs for local inhabitants. However, the effectiveness of these agreements varies, and their success hinges on persistent community engagement and oversight.
As AI infrastructure continues to expand, it increasingly factors into discussions about affordability. Escalating electricity demand, public incentives, and infrastructure investment alter the economic landscape families navigate. Connecting these dots reveals the profound impact of technological growth on living costs and highlights how policy decisions determine who truly bears those impacts.
Ultimately, the question narrows down to who truly benefits from the AI-driven economy. As Amartya Sen contends, “The central question is not what the system produces, but what it does to people.”
We shouldn’t allow AI to thrive at the expense of hardworking Americans’ finances.
