Leveraging AI

The boom in artificial intelligence has spawned a boom in data center construction—that is, building campuses of large warehouses whose (typically) 100,000 square meters of floorspace holds many floor-to-ceiling racks of advanced servers made up of computer chips, coolers and chilled water storage tanks, power distribution units and the audible hum of electricity underneath the much louder roar of cooling systems.  

These are not cheap; building a Tier IV data center will typically cost somewhere in the billions.  The data center on the new Apple Campus 2 will cost an estimated $5 billion.  And they’re getting bigger.  The Tahoe Reno 1 facility set the world record at 1.3 million square feet of floor space.  Not for long.  The Citadel Campus in Las Vegas, NV encompasses 7.75 million square feet.

At this point in the new construction boom, there are an estimated 5,420 data centers on U.S. soil, including 1,200 added last year.  Statistics differ on how many have come online this year because many are now being contracted in stealth locations, due to neighborhood resistance to power-hungry facilities that drive up electric bills for the entire community.  But it’s save to say that the boom is accelerating here and in Germany (529 facilities and growing), the UK (523), China (449), France (322) and Australia (314).  

Who’s paying for all this?  Corporate bond investors—and that could be a problem.

Big tech firms like Alphabet and Meta Platforms have been relying on huge bond offerings, expected to reach $1.5 trillion by 2028, to fund their AI data center construction, and the AI-related costs have recently reached 50% of all capital spending of all companies in the U.S. marketplace. 

But some bond analysts wonder whether these massive investments in artificial intelligence will pay off in the future.  A recent MIT report indicated that 95% of organizations are getting zero return from generative AI projects.  Stock analysts are also worried at the unprecedented amount of leverage in Big Tech firms.  If profits aren’t great enough to cover the debt obligations, bondholders could see a decline in the credit ratings of their holdings or, at an extreme, default—which could also trigger a decline in the share prices of the tech giants.

So far, this is speculation.  However, history tells us that not every competitor in a new market will survive the increasingly murderous scramble for market share.  All of the AI players expect to dominate the field and are investing as if that’s going to happen.  They won’t all be winners.

Sources:

• ‍https://www.eziblank.com/how-much-does-it-cost-to-build-a-data-center/ 
• ‍https://www.brookings.edu/articles/the-future-of-data-centers/ 
• ‍https://www.advisorperspectives.com/articles/2025/11/24/big-techs-ai-debt-wave-threatening-credit-markets
• ‍https://www.advisorperspectives.com/articles/2025/11/24/techs-debt-binge-raises-risk-race-create-ai-world