Against all odds, the US economy continues to grow at a record-setting pace. Yet as new economic data shows, that growth is anything but healthy, and predicated almost entirely on the AI boom.
A new essay by podcaster and writer Derek Thompson takes a look at key economic metrics surrounding AI in 2026. Among the most striking is a bar graph showing growth in construction spending, a key indicator of economic stability.
Using data from Harvard Kennedy School economist Jason Furman, Thompson shows that growth in data center construction has exploded this year, while all but one other category has regressed.
By the numbers, data center construction has grown at a rate of 64.5 percent in the most recent year, while new energy infrastructure has grown by 4.1 percent to keep pace — since demand is growing for electricity to power the AI boom.
Everything else has collapsed: multi-family home building fell by 5.6 percent, while healthcare and commercial retail construction both tumbled by 10.1 percent. New educational construction tumbled by 12.6 percent, while manufacturing — supposedly a key target for the Trump administration — has collapsed, shrinking nearly 24 percent from 2025 to 2026.
Thompson is perhaps best known for his “abundance” theory, a deregulation-brained policy agenda with roots in neoliberal economic thought.
But if there’s any abundance to be had in the US economy at the moment, it’s clearly coming on the back of AI — and with investors growing increasingly skeptical over the sustainability of AI spending, it could only be a matter of time before data center spending tumbles back down to Earth.
More on data centers: Retired Air Force Colonel Warns Americans That It’s Dangerous to Live Near Data Centers Because You Could Be Targeted or Killed