Resolved: The United States federal government should enact a moratorium on hyperscale data center construction.

What the resolution asks

A moratorium is a pause on new construction, not a ban on running existing facilities. Pro isn’t defending turning off anyone’s servers. It is defending stopping new hyperscale projects for some period while something else happens: grid planning, rules on who pays for new power, water rules, or a national siting standard. Strong Pro cases say what the pause is for and roughly how long it lasts, because an open-ended moratorium looks a lot like a ban and is much harder to defend.

"Hyperscale" has no single legal definition. Industry uses rough thresholds based on server count, floor area or power draw, and the largest campuses now need power on the scale of a small city. Expect definition fights. Pro wants a line that captures the giant AI campuses without touching ordinary enterprise data centers. Con can argue that any line pushes builders to split projects just under the threshold, so a moratorium on hyperscale buys less than Pro claims.

Where Pro wins: the grid and who pays

Pro’s strongest ground is electricity. Data center demand has become one of the main drivers of projected load growth in the US, and federal researchers at Lawrence Berkeley National Laboratory projected a steep rise in data center electricity use by the end of the decade. Utilities are planning new gas plants, delaying some fossil plant retirements, and building transmission to serve a small number of very large customers.

The ratepayer argument is where Pro gets its sharpest impact. When a utility builds generation and transmission for a data center, the question is who carries the cost if the demand forecast is wrong or the customer leaves. Some state regulators have approved special tariffs for large loads precisely because of this risk. Pro argues those fixes are piecemeal and that a national pause gives regulators time to set rules before households end up subsidizing the buildout through their power bills. Add water use for cooling in drought-prone regions and local opposition to siting, and Pro has a case about consent: communities are absorbing costs they never agreed to.

Where Con wins: competition, leakage and efficiency

Con’s headline argument is the AI race. Compute capacity is treated as a strategic asset, and federal policy in recent years has aimed to speed up data center permitting, not slow it. A national moratorium would be a sharp reversal, and Con argues rivals, especially China, would not pause with us. Weigh this carefully. It is a big impact with a long link chain, and Pro will ask whether a two-year pause on one category of building actually decides who leads in AI.

Con’s more technical arguments are often stronger. The first is leakage: the demand for computing doesn’t disappear, so construction moves to other countries or to smaller facilities below the hyperscale line. The second is efficiency. Hyperscale facilities are generally more energy efficient per unit of computing than the scattered smaller server rooms they replace, so a pause aimed at the most efficient builders could raise total energy use. The third is economic: jobs, local tax revenue, and the grid upgrades data center customers can help pay for. Counties like Loudoun in Virginia have built budgets around data center tax revenue, and Con can use that to argue the buildout can be managed rather than stopped.

The solvency fight

The round often turns on whether a pause solves anything. Con’s best line is that Pro’s problems (cost allocation, water rules, transmission planning) all have targeted fixes that don’t require stopping construction. If Con can show those fixes are already moving at the state and federal level, Pro’s moratorium looks like a blunt tool for a problem being solved with sharper ones. Pro’s answer is timing: the buildout is moving faster than regulators can write rules, and every project approved under today’s rules locks in costs for decades. A pause is how you make sure the targeted fixes arrive before the next wave of projects, not after.

International comparisons help both sides. Ireland’s grid operator effectively stopped new data center connections in the Dublin area because of capacity concerns. Pro can use that as evidence that pauses are a real policy tool. Con can use it as evidence that pauses push investment elsewhere. Know the details well enough to defend whichever reading you use.

How rounds get decided

Expect weighing between concrete, local and near-term harms (power bills, water, grid reliability) and large, national and speculative harms (losing the AI race, slower economic growth). Pro should make the AI race impact look speculative and the ratepayer harm look certain. Con should argue that the local harms are fixable without a pause, and that the competitive harm of pausing can’t be undone once rivals pull ahead. The team that wins the solvency debate usually gets to set those terms.