Elected officials right and left are cowering in face of news stories and angry town halls in opposition to data center construction. Even Texas, which has never met an economic development project it didn’t want, has paused approvals for new data centers at the direction of Gov. Greg Abbott.
A few stories, however, are about those projects lucky enough to have been completed before the furor erupted, and those lucky enough to be benefiting from them. For these fortunate few, a flood of new government revenue is offsetting property taxes, huge bonuses have been paid to teachers, and long-moribund communities are seeing new construction, revitalization and long desired growth.
But it won’t last forever. At some point, maybe sooner than you think, all that sweet, sweet data center investment will begin to dwindle, and some will have missed out.
The thing about buildouts, once they’re built out, they’re built out. Once the gold rush is over, it’s over. Or at least dwindles to a trickle. And no one seems to be taking that into consideration about data centers.
We tend to think trends will continue indefinitely, but they almost never do. Booms rise, crest, and then dwindle. There is every reason to think that the huge rush to build massive data centers will end at some point, and maybe sooner than you think. At least two reasons, in fact.
First, innovation in the AI space is happening at a breathtaking pace. The AI companies will find ways to design more compact models, and to train and run their models more efficiently. In fact, pretty soon AI will be redesigning itself to be faster, leaner and more efficient. One result may be less need for ever more and ever larger data centers.
And second, eventually Wall Street will get stingier with the money. Investors are already starting to ask harder questions about the return on all this investment and the billions they are putting at risk. How many data centers does a company need? How many AI companies are viable? As capital markets lose confidence in the return on investment, financing will tighten, and the data center buildout will slow accordingly. Some of these companies may consolidate, shrink, or disappear altogether. Anybody want a slightly used data center for 10 cents on the dollar?
Here’s the point: Put up barriers to data center development, and you might just miss out. And that would be unfortunate.
Data centers are terrific assets to a community or region. Compared with almost any other major commercial or industrial facility, data centers are clean, quiet, create high-paying jobs, generate little ongoing traffic and throw off considerable property taxes. They’re actually good neighbors.
But this pace of construction won’t last forever. Eventually the buildout will crest, capital will become more selective, technology will become more efficient, and the gold rush will dwindle. And if you missed out, you missed out, and you’ll probably regret it.