Artificial intelligence may live in the cloud, but the cloud has a very large physical footprint. Behind every AI query are enormous data centres consuming electricity, requiring cooling, occupying land and connecting to transmission systems. As governments race to capture the economic promise of AI, communities are beginning to ask a more basic question: who actually benefits economically—and who is left paying the costs?

In many cases, the imbalance is stark. The primary financial gains—data ownership, processing power, and platform value—flow back to Silicon Valley firms that own and operate the infrastructure. Local communities host the facilities, shoulder the strain on power and water systems, and often receive only a fraction of the long-term tax revenue in return.

Claim One: Data centres could place extraordinary new demands on electricity systems. — FACT.

The numbers are becoming difficult to ignore. Lawrence Berkeley National Laboratory now estimates that data centres could consume about 11.8% of all U.S. electricity by 2030, with plausible scenarios ranging from 9.5% to 15.3%. The U.S. Department of Energy says data centres are among the large new loads creating an urgent need for additional transmission infrastructure.

The concern for local communities is therefore legitimate. A data centre may arrive as a private investment, but if new generation, substations and transmission lines are required to serve it, those costs are often socialized—while the economic upside is largely captured by the firms that own the data and monetize the processing.

Claim Two: Data centres threaten community water supplies. — SOMETIMES FACT.

Cooling computers generates another demand: water. Berkeley researchers estimate U.S. data centres could directly consume roughly 140 to 280 billion litres of water annually by 2028, depending on how facilities are built and cooled. In Virginia—the world’s largest data-centre concentration—a 2026 state study warned that groundwater withdrawals, population growth and industrial demand could contribute to declining aquifers and recommended tighter controls.

But “data centres consume enormous quantities of water” is not universally true. New closed-loop and air-cooling technologies can dramatically reduce direct water use. Location and design matter. Still, even when environmental impacts are managed, the economic structure remains uneven: local regions absorb resource pressures while the value generated by computation and data processing is captured by global technology firms.

Claim Three: Data centres automatically bring enormous economic benefits to host communities. — PARTLY FICTION.

They certainly bring investment and substantial construction employment. Permanent employment, however, can be surprisingly modest relative to the billions invested and resources consumed. One enormous Kentucky development announced in July anticipates about 8,000 construction jobs but only 600 permanent positions.

That imbalance raises a central policy question: how much of the value created by data processing actually remains in the local economy? In many cases, communities receive a limited share through property taxes and negotiated incentives, while the far larger economic returns—cloud services, AI models, and data monetization—accrue to Silicon Valley companies that own the infrastructure and intellectual property.

Claim Four: The answer is simply to ban data centres. — FICTION.

That risks solving the wrong problem. AI capability increasingly depends upon access to computing infrastructure. Communities that simply prohibit data centres may protect themselves locally while pushing investment, innovation—and ultimately technological sovereignty—elsewhere.

The better question is under what conditions should they be permitted, and how should value be shared?

Alberta may already be pointing toward part of the answer. Provincial legislation encourages data centres to bring their own electricity generation and establishes the principle that data centres—not ordinary Alberta ratepayers—should pay for transmission upgrades they require. Alberta’s strategy also specifically identifies sustainable cooling, water availability and economic benefits for Albertans as considerations.

The Verdict: FACT—with an important qualification.

Data centres are becoming essential infrastructure for the future economy. But the current economic model often concentrates the benefits in global technology firms while distributing the costs across local communities that host them.

The emerging social contract should be straightforward: if Big Tech wants Big Data, it should bring—or fully pay for—the power infrastructure, conserve the water it uses, and ensure that a fair and meaningful share of the economic value from data processing and taxation remains in the communities that host it.

Otherwise, we may discover that the cloud was never really in the sky. It was sitting beside our communities all along—consuming resources that are locally borne while much of the economic reward flows elsewhere.