Cut out the middlemen who used to do it, and now use AI as the middleman.

By Cathy

In 2022, Galaxy Digital spent $65 million to buy a 180 MW Bitcoin mining farm. This year, the land started generating rent for CoreWeave under a 15-year lease, with average annual revenue of over $1 billion.

Crusoe was even more direct. It packaged and sold all 425 Bitcoin data center modules it had to NYDIG, then turned around and went to Texas to build Stargate for OpenAI.

Crypto.com spent $70 million on a domain name. The domain is AI.com, the most expensive domain transaction in public records. The product ran ads during the Super Bowl.

Three companies—every one of them is heavily pivoting to AI.

And this isn’t just another story of “AI plus blockchain.” They don’t issue tokens, don’t write whitepapers, and don’t build protocols. What they sell is compute power, server racks, and AI products. Their customers are OpenAI, Microsoft, and CoreWeave.

The money is moving in the same direction, too. In the first quarter of 2026, about 6,000 startups globally raised $300 billion, and AI took $242 billion—about 80%. In the same period, only 8 newly formed crypto funds raised a total of $1.1 billion, the fewest since Q3 2020.

If an industry that “eats tokens” can’t keep itself fed, why does it still get a slice in AI?

01 They’ve always been selling the same thing

First, look at Galaxy’s deal.

Helios is a mining site in West Texas, bought for $65 million in 2022. Galaxy stopped mining, invested $350 million of its own capital, and borrowed $1.4 billion in debt to renovate.

In the first stage delivered this year, 133 MW of critical IT load was sent to CoreWeave, and rent billing began. In the first three phases combined, CoreWeave committed to 526 MW under a 15-year lease, plus two additional five-year renewal options.

An asset bought for $65 million turns into a business that collects $1 billion in annual rent.

Crusoe’s path is even more complete. The company started out generating electricity from oilfield associated gas, and Bitcoin mining was just the first use for that power. In 2025, it sold its entire Bitcoin business—425 modular data centers, with none left behind.

The freed capacity went to Abilene, Texas. That’s OpenAI’s Stargate campus, planned at 1.2 GW. In May 2025, Crusoe raised $11.6 billion in debt and equity for it.

In March this year, Microsoft signed another 900 MW next door. The final Abilene plan totals 2.1 GW.

Worth noting is the timing. Galaxy bought Helios in 2024, and Tether also spent $420 million to buy Nvidia H100 in 2023.

Back then, ChatGPT had just caught fire, and AI data centers weren’t yet something everyone was fighting to get. These companies weren’t rushing into a new hype wave. They stood there first, and then the wind came.

At this point, the logic behind the pivot is actually pretty straightforward.

In the past decade, the one real skill crypto companies have learned is this: how to turn cheap electricity and a piece of land nobody wanted into a row of racks that can be powered, cooled, and run 24/7 at full capacity.

AI is exactly what’s missing right now.

So the group that already holds power and land can pivot with the least hassle. The assets are ready, the customers are new, and all that stands in between is a single round of renovation.

Others can pivot too, but they have to take a few more steps first. They need to prove that besides issuing tokens, they can do something else.

02 Not just selling land

Tether takes a different route.

Its AI division is called QVAC. What it does has nothing to do with token issuance: an open-source, cross-platform SDK that lets billion-parameter-class models run directly on laptops, regular GPUs, and even phones.

In March this year, QVAC released a LoRA fine-tuning framework for Microsoft’s BitNet 1-bit models; in April it released an SDK; in June it open-sourced Google Research’s VRAM compression algorithm TurboQuant. There’s also a local AI desktop app called Workbench.

A stablecoin company doing on-device inference optimization sounds absurd, but these things are real and on GitHub.

Tether’s money took an even more circuitous path. In December 2024, it made a strategic investment of $775 million in video platform Rumble. In June this year, Rumble completed its acquisition of Germany-listed Northern Data, obtaining 85.2% of the shares.

Northern Data operates Europe’s cloud business with more than 22k GPUs, plus data center capacity of about 250 MW by 2027. It raised its 2026 revenue guidance from €130–€150 million to €170–€190 million. A multi-year GPU cloud contract it signed is worth $270 million, and the customer is Together AI.

A stablecoin profit, routed around a video company, ultimately becomes GPU server rooms in Europe.

That $70 million from Crypto.com also wasn’t bought just to sit there. After Marszalek took over AI.com in April 2025, he started building a team. After the Super Bowl, the product launched: a personal AI agent that can send messages for users, execute actions across apps, and buy and sell stocks. He’s also CEO of two companies.

On the money end, Paradigm shut down the $1.2 billion 4th fund on July 8, explicitly writing AI and robotics. This crypto VC firm manages nearly $12 billion in assets and has already invested in delivery drone company Zipline and space defense company True Anomaly.

Most ironically, it’s OpenRouter.

Founder Alex Atallah was co-founder and CTO of OpenSea. He left in early 2023 and built something that works like a switchboard for more than 400 major models: developers call one API, and it decides whether the request goes to GPT, Claude, or some open-source model.

In May this year, OpenRouter raised $113 million in a Series B at a $1.3 billion valuation. 8 million registered developers, processing 2.5 quadrillion tokens per week.

A decade-long argument that the industry “doesn’t need middlemen”—the most successful company it produced is one that brokers for models.

Meanwhile, the money in crypto is draining out at a clearly visible speed. In April 2026, global crypto VCs invested only $659 million in a single month, the lowest in two years, down 75% year over year.

Why the money is leaving—Paradigm didn’t say much about it, but the LP ledger is clear. After FTX, Terra, and Three Arrows blew up one after another, institutional trust in crypto funds never recovered. In the last cycle, those high-valuation fundraising projects didn’t generate revenue and didn’t build real users.

At least, AI has bills you can check.

The last time this industry collectively changed direction was after the ICO collapse in 2018, when everyone pivoted to DeFi.

Back then, it was like swapping rooms in your own house.

This time, it’s moving out.

03 But the results are split

Not everyone pivoted into it.

Canaan is a counterexample. The company that makes Bitcoin mining machines also touched AI chips. In 2024, this business contributed about $0.9 million in revenue. In the same year, the company’s total revenue was $269.3 million.

$0.9 million versus $269.3 million. And this division consumed 15% of the company’s operating expenses.

In June 2025, Canaan shut down its AI chip business and went back to selling mining machines.

The mining companies are tiered too. Those with electricity, grid interconnection, and ready-made data center space landed multi-year contracts with Microsoft and Google. Those with only mining rigs and a press release—once the announcement is out, there’s nothing more.

The same 1 MW is a 15-year lease for a company that can handle delivery, but just a one-page slideshow for a company that can’t.

The distinction is really only one: is anyone actually paying for it?

Galaxy has CoreWeave’s rent. Crusoe has Microsoft and OpenAI’s campus. OpenRouter has 8 million developers calling its API. Canaan has a $0.9 million annual revenue figure—and then it figured out the numbers itself.

04 Summary

This group of companies isn’t doing “crypto AI,” they’re doing AI.

What’s really being transferred over isn’t technology. It’s a set of skills: how to raise billions when nobody believes you, and how to get power to a piece of wasteland within a dozen months. Those skills have little to do with blockchain—just happen to be trained in the crypto industry.

And it needs to be said clearly: moving out doesn’t mean winning. Crusoe’s Abilene campus for Microsoft won’t be powered until mid-2027. Crypto.com’s AI product just launched, and after Rumble finishes acquiring Northern Data, it still has to prove it can operate a data center.

None of these are confirmed until acceptance.

The only thing that’s certain is this: the bet has been placed, and it was placed on someone else’s table.

Ten years ago, they said they would eliminate the middlemen.

Now they’re becoming middlemen for AI.