The race to build AI infrastructure is the most important economic and national security issue of the next decade.

If the U.S. wants to remain the global leader in AI, we need to build data centers + the energy to power them.

One of the biggest arguments against new data centers is that they consume enormous amounts of water.

That’s NOT TRUE for the newest generation being developed today.

A typical U.S. 18-hole golf course uses roughly:

• 9–10 million gallons of water per month
• Around 300,000–370,000 gallons per day

Meanwhile, one recently approved 1.8 GW AI data center campus in Joliet, Illinois projects water consumption of only about 120,000 gallons per day.

That’s less than 40% of the average golf course!

And it’s possible because of technological innovation.

Modern AI campuses increasingly use closed-loop liquid cooling.

That means the water is:

• Filled once
• Continuously recirculated
• Only replaced for small leaks, maintenance, and occasional top-offs

Some of the newest closed-loop facilities project ongoing potable water use of only 5,000 gallons per day, largely for bathrooms and other domestic uses after the initial system fill.

What you actually get is an industry that generates:

• Thousands of construction jobs
• High-paying permanent technical jobs
• Significant tax revenue for local communities
• The computing infrastructure needed to compete globally in AI

That doesn’t mean every proposed project is appropriate or every concern is invalid. Local impacts on electricity, land use, and water availability should still be evaluated carefully.

But blanket opposition based on outdated assumptions about water consumption risks slowing one of America’s most important strategic industries.

If we fail to build the infrastructure, someone else will.