If There Is An AI Bubble, It Is Getting Bigger

Big Tech Is Spending More on AI Infrastructure — And Communities Are Pushing Back

Alphabet, Amazon, Meta, and Microsoft all reported earnings last week — and all four raised their capital spending plans, just one quarter into the year. At the start of April, their combined planned spend was approximately $670B. It now exceeds $725B (source: WSJ). The direction is clear: these companies are betting enormous sums on AI infrastructure, and they're accelerating, not pausing.

The Resource Constraint Nobody Is Talking About Enough

There is a lot to unpack here under the general umbrella of "we don't have unlimited resources." Resources limit the speed with which they can build data centers — with the chips (GPUs), memory, electricity, and water needed for cooling. At an Anthropic session today (5/5/2026), Andrew Ross Sorkin asked Jamie Dimon if $1 trillion was too much to invest in AI infrastructure this year. He basically said no — and added that the real uncertainty is in which companies will benefit.

Local Resistance Is Growing

The United States has more data centers than any other country. Virginia alone has 579. But communities are starting to say no.

Ten states have proposed legislation to block new data center construction. More than 90 local governments are doing the same. The objections are consistent: data centers generate tax revenue but create few lasting jobs — most employment ends when construction does. Beyond that, residents are worried about property values, environmental impact, and utility costs that are already rising faster than inflation.

One case is illustrative: a town of 14,000 is pushing back on a new facility despite an estimated $32M per year in local tax revenue. The money isn't the issue. What's driving resistance is something harder to address — fear of the unknown. Residents can't assess what the technology will mean for their water supply, their electric bills, or their daily lives. And they don't trust assurances from the government that it won't matter.

That trust gap is significant. The technology is moving faster than most people can process, and local communities — with little voice in decisions being made at a national or corporate level — are using whatever tools they have to slow it down.