In our last issue, I gave you one reason why the artificial intelligence boom could be headed for a painful correction.
The world’s largest technology companies are currently spending hundreds of billions of dollars building enormous data centers. But new technology could potentially change how much of that infrastructure we actually need.
And if that happens, some of the biggest investors in AI could lose fortunes.
But as Wall Street worries about whether we’re building too many data centers, the AI industry is facing a very different problem.
It still can’t build them fast enough.
America’s aging electrical grid is struggling to keep up with demand, and new projects can spend years waiting for power connections. Yet demand for AI computing power continues to grow.
So how do we keep expanding AI without relying entirely on massive, expensive data centers that are becoming increasingly controversial to build?
I believe the answer could change the way we think about AI.
And it could also create an entirely new generation of investment opportunities.
The Race to Find Power
Chipmakers keep finding ways to make AI processors faster and more powerful. But getting enough electricity to run them is another story.
Even existing data centers can only handle so many new chips before they run out of power or cooling capacity.
But that’s not the only obstacle the industry is facing.
Increasingly, Americans are pushing back against having enormous data centers built in their communities.
According to an ABC News/Washington Post/Ipsos poll released last month, 72% of Americans oppose having a new data center in their neighborhood.
And residents are doing a lot more than just complaining. During the first three months of 2026, local opposition blocked or delayed at least 75 data center projects worth approximately $130 billion.
And when you consider the cost of building these facilities, it’s easy to see why the industry would want to look for alternatives.
According to commercial real estate firm JLL, the average cost of constructing a data center is expected to reach $11.3 million per megawatt this year.
For a 100-megawatt facility, that’s more than $1 billion in construction costs alone.
What’s more, lenders are becoming more cautious about financing these expensive projects.
That’s why I believe one of the next big breakthroughs in AI could come from rethinking where we put all that computing power.
And we might already have part of the solution.
You see, America already has an enormous amount of electrical infrastructure in place. Power lines, transformers and electrical connections already serve millions of buildings across the country.
And while some parts of the grid are stretched to their limits, other locations have enough available power to run AI computers.
So rather than relying entirely on massive new facilities, what if companies could take advantage of some of those existing connections?
That’s the thinking behind a growing interest in distributed computing.
Instead of putting thousands of powerful chips in one enormous building, companies can spread certain AI workloads across smaller computing systems in different locations, all connected through the internet.
This approach has some important advantages.
For one thing, companies could potentially add computing power wherever suitable electricity and internet connections are already available, rather than waiting years for a single massive data center project.
And that could reduce the amount of new construction needed to expand computing capacity.
Plus, smaller installations should face fewer obstacles than the enormous facilities that have become so unpopular with local communities.
Of course, distributing computing power doesn’t eliminate the need for electricity or expensive hardware. And there’s still the question of whether the local power grid can support any additional demand.
But instead of building every new piece of AI infrastructure from the ground up, distributed computing would help companies make better use of what already exists.
As I explained in our last issue, Nvidia has been developing powerful AI systems that can operate outside traditional data centers. Meanwhile, manufacturers are producing modular computing facilities that can be deployed in smaller locations.
These systems won’t replace the enormous data centers that are needed to train the world’s most advanced AI models. But they could give the industry another way to expand its computing capacity as demand continues to grow.
And for investors, that raises an intriguing possibility.
Could some of the biggest opportunities in the next phase of the AI boom come from companies finding new ways to use the electrical infrastructure America already has?
That’s what I believe.
Because as valuable as Nvidia’s chips have become, they can’t do anything until somebody supplies the electricity to turn them on.
Here’s My Take
America needs more AI infrastructure, and I’m confident that enormous data centers will still play an essential role in the years ahead.
But as demand grows, getting access to electricity quickly could become just as valuable as having the latest AI chips. And companies that find ways to use existing infrastructure more efficiently could have an enormous advantage.
I believe some of the most exciting opportunities in the next phase of the AI revolution will come from technologies that make that possible.
Regards,
Ian King
Chief Strategist, Banyan Hill Publishing
Editor’s Note: We’d love to hear from you!
If you want to share your thoughts or suggestions about the Daily Disruptor, or if there are any specific topics you’d like us to cover, just send an email to dailydisruptor@banyanhill.com.
Don’t worry, we won’t reveal your full name in the event we publish a response. So feel free to comment away!
