I thought I’d answer this because last week we got a direct answer to this on Ciena’s earnings call, where Ciena’s CEO Gary Smith answered Ryan Koontz’s question about the political uncertainty of the construction pace of data centers directly, meaning the direct quote was “if they stopped building data centers tomorrow, you wouldn’t notice it for two years.”
The reason is that a significant portion of the demand is driven by GPU upgrades and the interconnects required for additional capacity in existing data centers, not just the construction of new buildings. The build-out is so massive that it’s hard to see it stalling out. And the necessary data centers will certainly be built somewhere.
In addition, last week X’s Global Government Affairs announced that during the summer, based on their suspicions, Chinese actors created approximately 200,000 X bot accounts aimed at influencing the construction of data centers, the discussions surrounding them, and perhaps political decisions and presumably also share prices through negative messaging campaigns.
https://x.com/GlobalAffairs/status/2093130747796148634
Based on the discussions I’ve followed, quite a few people have also changed their minds about data centers once it has been clearly demonstrated that the water consumption figures were previously way off, new closed/recycling systems are effective, property tax revenues are significant for municipalities, and the jobs are not just temporary construction work. There has been, and still is, a lot of misinformation prevalent.
The most vocal local groups have often been quite small and operating with very outdated information. To be honest, I wouldn’t want a data center next to my summer cabin either, and their mere presence in the neighborhood can cause a drop in property value. At the same time, however, I am more than happy to invest in data center companies. I agree with @Mustathmir that data centers will be built somewhere one way or another. There are plenty of municipal decision-makers eager to take them in.
As a final general and very condensed note, I finished a comprehensive deep-dive cross-analysis today of about 35 different AI layer quarterly reports and earnings transcripts, and the outlook for the AI Build-Out is excellent in 100% of the analyzed reports from 2027 all the way into 2028, and in some cases well beyond that. Financing risks naturally exist, but the demand is genuine and growing.