The $65 Million Bitcoin Mine That's Now Worth $1 Billion a Year

Galaxy Digital bought a struggling bitcoin mine in West Texas for $65 million in 2022. This week, they announced they're leasing it to CoreWeave for more than $1 billion a year.

Not because bitcoin rallied. Because they're evicting the miners and moving in AI training clusters instead.

Same building. Same substation. Different tenant. And suddenly, the asset just got repriced by 15x.

I've been watching crypto mining operations get revalued as AI infrastructure for months now, but two deals announced this week crystallized something I hadn't fully named: the most valuable thing bitcoin miners built wasn't the coin operation — it was the power contract.

The Repricing Nobody Saw Coming

TeraWulf just leased its entire Kentucky campus to Anthropic. Twenty years, roughly $19 billion, about 401 megawatts of AI compute capacity. Galaxy handed CoreWeave the first 133 megawatts at that West Texas site that used to belong to Argo Blockchain before they went under.

I read both press releases back to back. They tell the exact same story.

This isn't crypto companies pivoting to AI. This isn't diversification theater. This is the market discovering that what looked like speculative mining infrastructure was actually the scarcest asset in the AI race: permitted, interconnected, ready-to-energize power at scale.

And "ready" is doing a lot of work in that sentence. Building a new data center with serious power capacity doesn't take months. It takes years. Permitting, utility interconnection queues, substation upgrades, local opposition to industrial energy loads — I've watched clients burn eighteen months just getting in line for grid connection.

Bitcoin miners already did that work. They built in places nobody wanted them, negotiated interruptible power rates, upgraded substations to handle industrial loads, and then... the hashrate wars crushed their margins and AI showed up looking for exactly what they'd built.

The Fiber Parallel You've Seen Before

The telecom boom did this once.

In the late 90s, companies overbuilt fiber networks chasing the internet gold rush. Most of those companies went bankrupt. But the fiber they buried stayed in the ground, got bought for pennies on the dollar, and became the backbone that carried streaming video and cloud computing for the next two decades.

The freight changed. The infrastructure stayed valuable.

Nobody gets fired the day the fiber gets repriced. The bondholders just discover the asset they wrote off is suddenly worth something again.

That's what's happening now, except it's not fiber — it's substations, interconnection agreements, and permitted megawatts. The miners who survive this aren't the ones with the most efficient ASICs. They're the ones who figured out they're actually in the power business.

What Changes When the Asset Gets Relabeled

If you're a lender, this should make you rethink every crypto mining credit you've got on the books.

If you're an auditor, you're staring at companies whose primary asset just shifted categories — from "specialized crypto equipment" to "energy infrastructure with optionality." That's not a footnote. That's a different business.

If you're a finance leader trying to understand exposure, the question changed. You're not evaluating hashrate projections and bitcoin price sensitivity anymore. You're evaluating megawatts, interconnection queue position, and whether that site can realistically convert to GPU clusters.

I was on a call last month with a credit team trying to value a distressed miner. They kept modeling coin price scenarios. I asked: "What's the power contract worth if they never mine another block?" Long silence. They hadn't run that number.

They should. Because CoreWeave and Anthropic already did.

The Moat Was Always the Substation

Here's the uncomfortable part: most of the crypto industry spent the last five years arguing that bitcoin mining was securing a decentralized financial network. That the energy expenditure was justified by the trustless consensus mechanism. That the miners were infrastructure providers for a new monetary system.

All of that might still be true. But the market just said: "Cool story. We'll pay you more to train language models."

The asset that mattered was never the coin. It was the power.

And if you're a bitcoin maximalist, that's a hard sentence to read. It means the infrastructure you built to secure the network is now worth more doing something else. It means the energy you argued was essential to bitcoin's security model is getting bid away by whoever needs compute capacity and can't wait three years for a new data center.

Which raises a question I'm not sure anyone has a clean answer to: if the most valuable mining infrastructure exits to AI over the next 24 months, what happens to bitcoin's hashrate? Does it drop? Does it stabilize at a lower equilibrium? Do new miners step in with cheaper power in less convenient locations?

I don't know. But I know the incentive structure just changed.

Who Else Is Mispriced?

So here's what I'd be asking if I were sitting in your seat Monday morning:

If you have exposure to crypto mining companies — as a lender, investor, auditor, or advisor — pull the list. Not the hashrate. Not the coin holdings. Pull the power contracts. Where are they interconnected? How many megawatts? What's the queue position if they wanted to add capacity? Is the site zoned and permitted for data center conversion?

Because those aren't crypto questions anymore. Those are energy infrastructure questions. And if you're still valuing those companies like they're leveraged bets on bitcoin, you're missing what the buyers are actually pricing.

Which names on your watchlist are labeled "crypto" but are really power companies you're undervaluing?

Galaxy bought that Texas site for $65 million thirty months ago. The contracted revenue CoreWeave is paying suggests it's worth something closer to $10-15 billion over the life of the deal. That's not a rounding error. That's a category change.

The railroad didn't make the town valuable. The railroad revealed which towns were sitting on valuable land. Same thing here. Bitcoin mining didn't create the power infrastructure's value. It just forced someone to build it in the middle of nowhere, right before AI showed up willing to pay 15x for the keys.

I've watched this movie four times now — telecom fiber, colo data centers, content delivery networks, and now mining infrastructure. The platform shifts. The infrastructure gets repriced. The companies that survive are the ones who realize they were in a different business than they thought.

If you're waiting for clarity before you revalue these assets, you're already behind. The deals are getting signed. The power is getting re-tenanted. The only question is whether you repriced it before or after everyone else figured it out.

---

What to do this week: Pull your portfolio exposure to anything labeled "crypto mining." Ask your team: do we know the megawatts, the interconnection status, and the conversion optionality? If the answer is no, you're flying blind. If the answer is yes, you might be sitting on an energy infrastructure play that your models are still treating like a speculative bet on coin price.

Either way, the asset just got relabeled. Time to revalue it.