When Isabelle Reksopuro heard Google was taking public land for data centers in Oregon, she couldn't sort fact from fiction. So she built an interactive map to track data center construction and AI policy.
The reality turned out more complicated than the headlines suggested. The Dalles, a city of about 16,000 near the Washington border, requested 150 acres of Mount Hood National Forest for municipal water needs. Google isn't technically claiming the land, they're just the massive power user driving the demand.
This is the misinformation problem Reksopuro wanted to solve. Data centers are expanding fast to support AI infrastructure, but the details of who's building what and where get murky in local news and social media.
Her map gives people a way to see what's actually being built in their area. As AI companies race to add compute capacity, these facilities are popping up in communities that often don't realize what's coming until construction starts.
For anyone working with AI tools, this matters because your models run somewhere physical. The infrastructure buildout happening right now will determine which companies can scale, how much compute costs, and whether your favorite AI service stays fast or slows down when everyone else is using it too.
