The Quiet Town That Said No to Big Tech’s Hunger for Power
In the shadow of New Jersey’s Turnpike, where warehouses and strip malls blur into a landscape of logistical convenience, a proposed data center in New Brunswick became a flashpoint in a national reckoning. Local officials rejected the project not over aesthetics or traffic, but over a more fundamental concern: the insatiable energy demand of the digital infrastructure powering artificial intelligence. The decision, while seemingly local, reflects a growing tension between municipalities and the tech industry’s unchecked expansion—a tension that could reshape how and where the next generation of AI is built.
Data centers are the invisible backbone of the modern internet, but as AI models grow more complex, their power consumption has skyrocketed. A single large language model training run can consume as much electricity as 1,000 U.S. homes use in a year. The New Brunswick proposal, backed by a major cloud infrastructure provider, promised jobs and tax revenue but required a dedicated substation and would have drawn enough power to supply a small city. For a town already grappling with aging infrastructure and rising energy costs, the math didn’t add up.
Energy, Not NIMBYism, Drove the Decision
Contrary to the narrative that local opposition is rooted in resistance to progress, the rejection was grounded in pragmatic concerns about grid reliability and long-term sustainability. New Brunswick’s planning board cited insufficient evidence that the local utility could handle the additional load without risking brownouts or forcing upgrades paid for by ratepayers. The developer offered financial incentives, but failed to present a clear plan for energy sourcing beyond tapping into an already strained regional grid.
This isn’t about blocking technology—it’s about accountability. As AI workloads multiply, data centers are no longer passive facilities. They are dynamic, power-hungry operations that demand real-time energy at scales once reserved for heavy industry. The New Brunswick case underscores a critical flaw in current development models: the assumption that energy is an infinite commodity, and that communities will absorb the externalities of tech’s growth.
Other towns are watching. In Loudoun County, Virginia—home to the world’s largest concentration of data centers—residents have raised alarms over rising electricity bills and environmental degradation. In Oregon, regulators delayed permits for new facilities until developers could prove they wouldn’t exacerbate water shortages. These aren’t isolated incidents. They’re symptoms of a systemic imbalance: the tech industry’s pace of innovation outstripping the physical and civic infrastructure needed to support it.
The AI Gold Rush Meets Physical Limits
The denial in New Brunswick arrives at a pivotal moment. Tech giants are racing to deploy AI at scale, and every new model demands more computing power, which in turn demands more data centers. But unlike software, data centers can’t be spun up in the cloud—they require land, water, and megawatts of electricity. The industry’s growth is hitting the hard limits of the physical world.
Consider the numbers. Global data center electricity use is projected to double by 2026, driven largely by AI. In the U.S., data centers already consume about 2% of total electricity—more than some entire countries. Yet, permitting processes haven’t evolved to account for this shift. Most local governments lack the technical expertise to evaluate energy impact assessments or negotiate long-term utility agreements. The result is a patchwork of approvals based on outdated assumptions, often favoring short-term economic gains over long-term resilience.
New Brunswick’s decision signals a shift toward more rigorous scrutiny. It’s not that the town is anti-tech; it’s that it’s demanding a seat at the table. Developers must now prove not just that a data center can be built, but that it can be powered responsibly—without burdening residents or destabilizing the grid. This is a reasonable ask, and one that should become the standard.
A Blueprint for Smarter Growth
The rejection isn’t a setback for AI—it’s a necessary correction. The industry can no longer treat energy as an afterthought. Forward-thinking companies are already exploring alternatives: co-locating data centers with renewable energy sources, investing in on-site generation, or designing facilities that reuse waste heat. Microsoft’s underwater data center experiment and Google’s push for 24/7 carbon-free energy are steps in the right direction, but they remain exceptions.
What’s needed is a new framework for data center development—one that integrates energy planning, community input, and environmental impact from day one. This means transparent modeling of power demand, binding commitments to renewable sourcing, and shared investment in grid upgrades. It also means rethinking location strategy. Instead of clustering in already saturated regions like Northern Virginia or Silicon Valley, developers should consider distributed models that leverage underutilized infrastructure in smaller markets, where energy capacity may be more readily available.
New Brunswick may have said no to one project, but its message is clear: the era of unchecked expansion is over. As AI reshapes industries, it must also adapt to the realities of the physical world. The companies that thrive will be those that treat energy not as a cost center, but as a core design constraint—and that engage communities not as obstacles, but as partners.
The future of AI isn’t just about algorithms and processing power. It’s about whether we can build the infrastructure to support it—without leaving towns like New Brunswick to foot the bill.