Loudoun County, Virginia, is home to around 250 data centres, the highest concentration in the world. In 2024, the revenue they generated exceeded the county's operating budget by 35 million dollars. The property tax rate has fallen from around $1.29 to 80 cents per $100 of value since 2008, and twenty-two schools have been built in fifteen years. Residents who live next to the facilities no longer want them.
We have documented local opposition in several forms, from India to Pennsylvania. This case is different, and it is more instructive, because it prevents us from telling a simple story.
The narrative this case contradicts
Data centres are often presented as facilities that impose their nuisances without giving anything back to the area. Loudoun demonstrates the opposite.
Revenue higher than the county's operating budget. A fall of more than a third in the property tax rate. Schools built. For a local taxpayer, the benefit is concrete, quantified and visible every year on their tax bill.
This is the argument operators make, and in Loudoun it checks out. It needs to be said clearly before we turn to the rest.
The other narrative it also contradicts
Conversely, these tax windfalls are presented as proof that the development benefits everyone. Loudoun demonstrates that this is not the case either.
The tax benefit is spread across the whole county. The nuisances are concentrated on those who live nearby: the constant noise of cooling systems, the high-voltage power lines, the transformation of the landscape, the strain on the electricity grid.
It is the classic pattern of facilities that everyone benefits from and nobody wants next door. It cannot be resolved by showing that the overall balance is positive, because the people who pay and the people who gain are not the same.
Property values in the county are said to have risen by as much as 182% in ten years. One might see this as proof that the presence of the facilities has done no harm. That would be to confuse an average with a distribution. The rise benefits the county as a whole, buoyed by its prosperity; it says nothing about the houses at the foot of a cooling building, whose relative value may have moved quite differently. This is exactly the problem we raised about indicators of progress: a positive aggregate can mask concentrated losses.
What this case teaches us
Loudoun is ahead of the curve. What is playing out there today will play out elsewhere tomorrow, and three lessons emerge.
The tax argument works, then runs out of steam. It wins support as long as the facilities remain few in number. Beyond a threshold of density, the accumulated nuisances exceed what the diffuse benefit can offset in the eyes of nearby residents.
Compensation must be targeted. A benefit spread across an entire region does not offset a nuisance concentrated in a few neighbourhoods. The mechanisms that work return a share of the revenue to those who bear the direct impact.
The question is shifting to electricity. That is what the 417 to 3 vote confirmed last week: making installations pay for the grid reinforcements they cause addresses exactly one of the residents' complaints.
Key takeaways
This case is worth knowing about because it disarms both sides. Supporters can no longer claim the nuisances are imaginary. Opponents can no longer claim the areas gain nothing from it.
The real question it raises is neither technical nor ideological: it is distributive. Who receives the benefit, who bears the cost, and is there a mechanism that brings the two closer together?
This is probably the most useful way to frame the debate everywhere else. Not "do we need data centres", a question whose answer will be yes, but "who pays and who cashes in", a question that can be given concrete answers.