Technology companies want them. Governments want them. Real estate developers want them. Utilities want them and the business that comes with them.
Investors want a piece of whatever company can supply the land, electricity, chips, cooling systems, cables, transformers, and concrete needed to build them.
And the numbers are getting ridiculous…
Trillions of dollars are expected to flow into artificial intelligence infrastructure over the next several years.
Companies are buying land, signing long-term power agreements, restarting old power plants, considering nuclear reactors, and searching for any location where they can secure enough electricity to run the next generation of computing.
The entire world seems to be asking the same question:
Where are we going to build all these data centers?
But I think that’s the wrong question. Or at least an incomplete one…
Because while everyone is searching for places to build the data centers of the future, there are already enormous computing facilities scattered around the world.
They already have land. They already have power. They already have cooling. They already have industrial infrastructure.
They’re often located in politically stable countries with relatively predictable laws and property rights.
And many of them were specifically built in places where electricity was cheap and plentiful.
The problem is that a lot of them are being used to mine cryptocurrency, but that might be about to change…
The Original Search for Cheap Power
Long before artificial intelligence companies started scouring the planet for electricity, cryptocurrency miners were already doing it.
That was their entire business model…
Bitcoin mining is basically a competition to perform enormous amounts of computation as efficiently as possible where the winners are rewarded with newly created Bitcoin.
That means the economics depend heavily on the cost of electricity.
So miners went searching for the cheapest power they could find…
They moved near hydroelectric dams. They set up operations in areas with abundant natural gas.
They moved into cold climates where outside air could help reduce cooling costs.
They bought old industrial sites with existing electrical infrastructure. They negotiated directly with utilities.
They looked for locations where they could connect enormous amounts of computing equipment to the grid without waiting years for approval.
Sound familiar? Well, it should…
Because that’s almost exactly what AI companies are doing today.
The difference is that the cryptocurrency miners started almost a decade earlier.
The Hardest Part Is Already Done
When most people picture a data center, they picture a giant warehouse filled with computers.
But the building isn’t the hard part. You can build a warehouse almost anywhere.
The real problem is everything that has to come with it…
You need enormous amounts of reliable electricity. You need grid connections. You need substations. You need transformers. You need transmission capacity.
You need cooling. You need fiber. You need roads. You need security. You need permits. You need local governments willing to let you operate.
And, increasingly, you need time. And that might be the most valuable asset of all…
You see, the AI industry doesn’t have ten years to wait for every new power plant, transmission line, and data center campus to work its way through the permitting process.
The race is happening now.
Companies that can get computing capacity online in 18 months could have an enormous advantage over companies that have to wait five years.
And that changes the value of existing infrastructure…
A cryptocurrency mining site might not be ready to host advanced AI processors tomorrow morning. To be fair, in many cases, it will need substantial upgrades.
AI servers are more power-dense. Their cooling requirements can be more complicated. The networking infrastructure is different. Reliability standards can be higher.
But that misses the bigger point…
The building can be upgraded. The servers can be replaced. The cooling systems can be improved…
But the electrical connection is much harder to recreate. And in the AI age, access to electricity may end up being worth more than the computers plugged into it.
Yesterday’s Crypto Mine Could Be Tomorrow’s AI Factory
This is where the investment opportunity gets interesting…
The market spent years valuing cryptocurrency miners based primarily on a handful of numbers.
How much Bitcoin could they produce? What did it cost them to produce it?
How much computing power did they control? How much Bitcoin did they hold on their balance sheets?
Those numbers still matter. But they may no longer tell the whole story…
Because some of these companies aren’t just cryptocurrency miners.
They’re owners of powered land. They’re owners of grid connections.
They’re owners of data center campuses. They’re owners of cooling infrastructure.
They’re owners of something the AI industry desperately needs and can’t quickly manufacture: Time.
And that means a cryptocurrency mining company that looks mediocre based on its Bitcoin production could be sitting on an extremely valuable collection of infrastructure.
The market may think it owns a struggling crypto mine, but an AI company may look at the exact same property and see a shortcut worth hundreds of millions of dollars.
That difference in perception is where investors should be looking.
The Great Compute Migration Has Already Started
This isn’t just a theory anymore…
Cryptocurrency miners have already begun shifting parts of their businesses toward artificial intelligence and high-performance computing.
The reason is simple: Mining cryptocurrency can be a volatile business.
Revenue depends on cryptocurrency prices, network competition, equipment efficiency, and electricity costs.
A miner can spend enormous amounts of money building infrastructure only to watch the economics change when crypto prices fall or competition increases.
AI offers a different potential business model…
Instead of using all that electricity to compete for digital coins, a company may be able to lease infrastructure to AI developers, cloud providers, or other high-performance computing customers.
That can create longer-term contracts and more predictable revenue. And investors are beginning to notice.
Wall Street is increasingly looking at certain cryptocurrency miners not as pure crypto companies…
But as potential AI infrastructure companies with something extremely valuable already in place: power.
That doesn’t mean every cryptocurrency mine will become an AI data center.
Some are in the wrong locations. Some don’t have adequate fiber connections.
Some don’t have the right buildings. Some use power arrangements that aren’t reliable enough for AI workloads.
Some were built as cheaply as possible and would require too much money to upgrade.
But some of them are sitting on exactly what the AI industry needs. And those are the ones worth watching.
Stop Counting Servers and Start Counting Megawatts
For years, investors looked at cryptocurrency miners and focused on computing power. But I’m convinced the more important number today is electrical power…
How many megawatts does the company already control? How much additional power can it secure?
How long are its energy agreements? What does that power cost?
Is it connected to a reliable grid? Can the site be expanded?
Is there fiber nearby? Can the infrastructure support high-density computing?
And how quickly could the site be converted?
Those questions can tell you something a Bitcoin production chart can’t…
They can tell you what the company might become.
And that’s important because the next generation of winners may not be the companies that mine the most cryptocurrency.
They may be the companies that realize their real asset was never the cryptocurrency mine in the first place.
It was the infrastructure underneath it.
The Picks and Shovels Are Already in the Ground
Investors love a good gold rush.
But the biggest winners are often the people who already own the things everyone suddenly needs.
During the AI boom, everyone is searching for chips. They’re searching for copper.
They’re searching for transformers. They’re searching for land. And, more than anything else, they’re searching for electricity.
Cryptocurrency miners spent the last decade searching for many of the same things.
They found cheap power. They secured land. They built server farms. They installed cooling systems. They connected to the grid.
They did all of this because they thought the future would belong to cryptocurrency, but some of them may have accidentally built infrastructure for an even bigger market.
That’s the opportunity investors need to understand…
The AI data center boom won’t be built entirely from scratch. It can’t be. There isn’t enough time.
Instead, some of the most valuable infrastructure in the AI race may already be sitting in plain sight, hidden inside companies that investors still think of as cryptocurrency miners.
The machines inside those buildings can change. The chips can change. The customers can change. The business model can change.
But cheap, reliable power connected to a functioning data center campus is much harder to replace.
That’s why, as the AI race accelerates, I’m spending less time asking who can build the biggest new data center.
I’m starting to ask a different question: Who already owns one?