bilateral
TopicsAbout← Feed
BySRSam Reyes·CMCal Morrow·EQEliza Quinn·DPDana Park
BREAKINGMay 8, 2026

In Defense of Data Centers

A growing political and policy debate has emerged over data center expansion across the United States, with politicians from both parties pushing bills to restrict new development and some proposing a nationwide moratorium. Defenders of the industry argue the opposition is rooted in short-termism and misinformation, while critics cite concerns over rising electricity rates, water consumption, and rapid land use changes. The debate has intensified alongside the AI boom, which is dramatically accelerating demand for new data center capacity.

XLinkedInFacebookThreadsWhatsAppReddit
⚡Both sides have significant internal splits on this story. Arguments below represent the dominant positions — see The Divide below for the full picture.

Data centers consume enormous power and water while generating minimal local jobs. Who bears the real cost of the tech industry's infrastructure, and who actually benefits?

Not familiar with this story? Get context →
Jobs promise versus jobs delivered
Liberal
The most telling number in this debate isn't $727 billion — it's 20 to 50. That's the on-site headcount of a typical operational data center. Microsoft or Amazon secures hundreds of acres, locks in preferential utility rates, draws enormous power from a shared grid, and then employs roughly the same number of people as a mid-sized Applebee's. The promise communities are sold is manufacturing-era prosperity. What they get is a server farm that runs itself.
Conservative
The Applebee's comparison is vivid, but it proves too much — a water treatment plant employs about 12 people and we don't call that extraction. The question isn't headcount; it's what the facility produces for the community around it. You've already acknowledged the 6.5-job multiplier and Loudoun's schools and tax cuts — those are real people with real paychecks, not a rhetorical device.
Liberal
Loudoun is thirty years of cumulative density — it's not what the next facility in a rural Virginia county gets to expect. And that 6.5 multiplier comes from an industry-aligned study whose methodology hasn't been independently verified; we shouldn't restructure local power grids around a number that may have been produced specifically to win this argument.
Conservative
Fair to flag the sourcing, but the burden runs both ways — if you're asking communities to refuse development on the basis that the multiplier might be inflated, you need independent evidence it's wrong, not just that it's industry-funded.
Loudoun County as replicable model
Liberal
Loudoun County funds nearly half its government budget from data center tax revenue — that's a real balance sheet, and I won't minimize it. But Loudoun is the exception engineered by thirty years of cumulative data center density, favorable geography, and proximity to federal demand. Treating it as the template for 3,000 new facilities spread across less affluent, less established jurisdictions is like pointing to Houston's energy wealth to argue that every county with an oil well will fund its schools.
Conservative
The Houston analogy actually helps my case: Houston didn't become an energy hub by accident — it got there through consistent permitting, infrastructure investment, and not calling moratoriums every time a refinery was proposed. The path to more Loudoun Counties is more data centers, not fewer. You're arguing that because a model hasn't yet replicated everywhere, we should restrict the mechanism that produces it.
Liberal
The difference is that Houston's wealth came with jobs that actually stayed in Houston. What's replicating across jurisdictions right now isn't Loudoun's fiscal windfall — it's the power draw, the rezoned farmland, and the 40-person headcount.
Conservative
Then the argument is for better community benefit agreements in those specific jurisdictions — not a policy that prevents any jurisdiction from becoming the next Loudoun before it gets the chance.
National water statistics obscuring local harm
Liberal
The 0.05 percent national freshwater figure is doing enormous rhetorical work while concealing the actual impact. A single hyperscale campus in drought-stressed Arizona can draw millions of gallons per day from aquifers that local agriculture and residents depend on — the national percentage tells you nothing about whether a specific community's water table survives the next decade. This is the same statistical sleight-of-hand the chemical industry used for generations: nationally, our emissions are a fraction of the total; locally, your children's asthma rates are irrelevant to the aggregate.
Conservative
You've described a permitting failure, not a data center problem. If a facility is drawing unsustainable water from a stressed aquifer, the correct fix is hydrology-based permitting standards that block that specific siting — not a national policy triggered by the aggregate figure you just said proves nothing about local conditions. The chemical industry comparison cuts against you: we didn't ban chemistry, we built the EPA and NEPA.
Liberal
Right — and NEPA reviews for data centers are routinely expedited or waived under economic pressure, which is precisely why 'trust the permitting process' rings hollow when the permitting process is the site of the problem.
Conservative
Then fix the expediting, not the industry. A moratorium that stops projects in water-rich Pennsylvania because of Arizona's aquifer stress is the bluntest possible instrument for a problem you've correctly diagnosed as localized.
Who pays the utility cost transfer
Liberal
Electricity rate increases following major data center siting land hardest on residential ratepayers who had no vote in the permitting process. That's a transfer of cost from shareholders of trillion-dollar corporations to families paying utility bills in communities that were promised a windfall. The fiscal benefit flows to the county government's balance sheet; the rate increase hits the household budget directly.
Conservative
Utility rate structures are set by public utility commissions, not by data center developers — if large industrial users are being cross-subsidized by residential ratepayers, that's a regulatory failure you should want to fix at the commission level. Loudoun County families paying lower property taxes because of data center revenue are also households with budgets. You're counting the rate increase and not counting the tax cut.
Liberal
The tax cut goes to whoever owns property; the rate increase goes to whoever uses electricity, including renters who own nothing and received no windfall. Those aren't the same households.
Conservative
That's a real distributional point — and the answer is rate design reform that charges industrial users their actual infrastructure costs, which I'd support. It's still not an argument against the data center; it's an argument against bad utility regulation.
National security versus democratic accountability
Liberal
The AI infrastructure race is real — I'll say so plainly. But 'China is building data centers' is not a sufficient answer to 'who pays when our water bills double and our grid strains every summer.' The U.S. built the interstate highway system with genuine community input and land-use law. The idea that we can't build AI infrastructure with similar democratic accountability is a corporate preference dressed up as a national security argument.
Conservative
The interstate highway system displaced hundreds of thousands of people with minimal compensation and bulldozed entire urban neighborhoods — it's an odd model to invoke for democratic accountability. And the arms-race framing isn't corporate preference: compute capacity is the proximate constraint on AI development right now, and construction pipelines take years to restart once interrupted. A moratorium doesn't pause the race; it hands the lead to an opponent who isn't pausing.
Liberal
Granting that compute is a real constraint, the question is whether the cost of building faster falls on shareholders or on the families living next to the facility. Speed and democratic accountability aren't mutually exclusive — they require political will, not a blank check.
Conservative
Binding community benefit agreements, transparent utility impact disclosures, hydrology-based permitting — I've agreed to all of it. The remaining disagreement is whether those tools require slowing deployment or just governing it better, and I think the evidence points to the latter.
Conservative's hardest question
The on-site employment figure — typically 20 to 50 workers per facility — is a genuine vulnerability. Communities that rezone land, absorb grid strain, and accept rapid development have a reasonable expectation of local workforce benefit, and the multiplier figures that justify the jobs case are industry-commissioned and not independently verified. Conservatives who defend this industry honestly must reckon with the fact that some of the communities absorbing the costs are not capturing proportionate benefits, and 'the broader economy gains' is cold comfort to a town that rezoned its farmland for a building staffed by three security guards and a network engineer.
Liberal's hardest question
The most difficult challenge to my argument is that Loudoun County's fiscal model — however exceptional — is genuinely real: schools funded, taxes cut, public services sustained. If even a fraction of the 3,000 planned facilities produce comparable local fiscal benefits, the harm-versus-benefit calculus becomes harder to sustain without stronger empirical evidence of systematic community damage rather than anecdotal cases.
The Divide
*The data center boom splits both parties between national competitiveness and local control.*
TECH-RIGHT
Data centers are essential national infrastructure that should be built rapidly with minimal regulatory obstruction.
“Halting AI projects would be a profound mistake.” — National Review
LOCALIST
Communities deserve meaningful control over data center development to protect local resources, grid stability, and character of place.
PROGRESSIVE LEFT
Data center expansion for AI is a corporate land grab requiring strict environmental review, moratoriums, and community veto power.
“The dark side of the data center boom.” — CODEPINK
MODERATE DEMOCRAT
Data centers bring fiscal benefits and should be welcomed with negotiated community agreements addressing local concerns.
The Verdict
Both sides agree
Both sides accept that data centers concentrated in Loudoun County have generated genuine fiscal benefits—schools funded, taxes cut—proving the economic model works in at least some contexts, which means the policy debate is really about replicability and distribution of costs, not whether the industry produces value.
The real conflict
FACTUAL: Whether the 6.5-jobs-per-worker multiplier and $727 billion GDP figures are reliable enough to guide policy—the conservative accepts them as establishing baseline economic case; the liberal treats them as industry-produced numbers whose methodology is unknown and therefore unsuitable for restructuring local infrastructure around them.
What nobody has answered
If data centers are genuinely replicable fiscal wins for communities like Loudoun County, why have the 3,000 planned facilities not yet signed the kinds of binding community benefit agreements and transparent utility-impact disclosures that would prove the model works elsewhere—and does the absence of such agreements suggest that developers see the Loudoun exception as non-replicable rather than non-negotiated?
Sources

More debates