David Sacks Says Data Centers Cut Property Taxes by $16,000 in Virginia County
Sacks offered the Virginia and Louisiana cases as examples of local benefits, without county budget data to check them.

All-In with Chamath, Jason, Sacks & Friedberg · GPT-6 Hits AGI? Tech Euphoria 2.0, SF Mansion Shortage, NYC Bans AI in Schools & Venezuela Oil DealDavid Sacks said revenue from data centers cut the average taxpayer's property tax bill by $16,000 in Loudoun County, Virginia, offering his figure as an example of a good local deal.
Sacks, who said he worked on a December executive order on artificial intelligence, argued the decision should stay local and practical, with communities negotiating terms in which companies stand up net new power generation, pay for grid upgrades, and can lower electricity costs.
He also cited a Louisiana county he did not name, where he said data-center revenue paid for teachers' first bonus in years.
He said concerns about heavy water use were overstated because the systems recirculate, and that centers can be placed where they do not bother people.
Questions people ask
How can a data center lower my property taxes?
Local governments tax the buildings, servers, and equipment inside data centers as commercial property. That adds a large new source of revenue, which can let a county lower the tax rate for homeowners while still funding schools and services.
Do data centers raise electricity bills for neighbors?
They can, but it depends on the deal. A large data center adds major demand to the local grid, and if the utility builds new capacity and passes the cost to everyone, household rates can rise. Many recent agreements require the operator to pay for new generation and grid upgrades to avoid that.
Do data centers use a lot of water?
Modern facilities often use closed loop systems that recirculate water for cooling, so they withdraw far less than older designs. They still use water, especially in hot weather, and the exact amount varies by location, design, and whether they use air cooling.