Multiple crypto companies are transforming into AI relying on the advantages of power plants, capital is flowing into AI, and the industry is shifting from intermediaries to AI intermediaries.

Written by: Cathy

Galaxy Digital spent $65 million to acquire an 180 MW Bitcoin mine in 2022. This year, this land began renting to CoreWeave with a 15-year lease, generating more than $1 billion in annual income.

Crusoe is even more straightforward. It sold its 425 Bitcoin data centers as a whole to NYDIG, then turned to Texas to build Stargate for OpenAI.

Crypto.com spent $70 million to acquire a domain name. The domain is called AI.com, the most expensive domain transaction in public records. The product was advertised during the Super Bowl.

All three companies are strongly shifting towards AI.

This is not another round of "AI plus blockchain" story. They are not issuing tokens, writing white papers, or creating protocols. What they sell is computing power, cabinets, and AI products, with clients like OpenAI, Microsoft, and CoreWeave.

Money is also heading in the same direction. In the first quarter of 2026, around 6,000 startups globally raised $300 billion, with AI taking $242 billion of that, accounting for 80%. During the same period, only 8 new crypto funds were established, totaling $1.1 billion, the lowest since the third quarter of 2020.

How can an industry that relies on tokens secure a position in AI?

01 They have always sold the same thing

Let’s first look at Galaxy's account.

Helios is a mine in West Texas, purchased for $65 million in 2022. Galaxy stopped mining and invested $350 million of its own funds, borrowing an additional $1.4 billion to make renovations.

In the first phase this year, a 133 MW key IT load was delivered to CoreWeave, and rent calculations began. In total, CoreWeave has committed to 526 MW across three phases, a 15-year lease, along with two five-year renewal options.

A $65 million acquired asset became a business generating $1 billion in rent annually.

Crusoe's path is even more radical. This company started with power generation from associated gas in oil fields, and Bitcoin mining was just the first use of that electricity. By 2025, it sold its entire Bitcoin business, 425 modular data centers, leaving none behind.

The freed-up hands were used in Abilene, Texas. There is OpenAI's Stargate No. 1 campus, planned for 1.2 GW; in May 2025, Crusoe raised $11.6 billion in debt and equity for it.

This March, Microsoft signed for another 900 MW next door. The entire Abilene is ultimately planned for 2.1 GW.

It's noteworthy to observe the timing. Galaxy bought Helios in 2024, while Tether spent $420 million on Nvidia's H100 also in 2023.

At that time, ChatGPT had just gained popularity, and AI data centers were not yet something everyone was scrambling for. These companies did not chase the hype; they were already standing there when the wind blew in.

At this point, the rationale for transformation is actually very simple.

What crypto companies have truly learned in the past decade is only one thing: how to turn cheap electricity and an unwanted piece of land into a row of cabinets that can be powered, heat dissipation managed, and fully utilized around the clock.

What AI currently lacks is exactly that.

Therefore, those holding electricity and land have the easiest transition. The assets are ready, the clients are new, and only a single renovation separates them.

Others are not unable to transform; they just need to take a few more steps. They must first prove that beyond issuing tokens, they can do something else.

02 Not just selling land

Tether is taking a different route.

Its AI department is called QVAC, and what it does has nothing to do with issuing tokens: a set of open-source cross-platform SDKs that allow billion-parameter models to run directly on laptops, regular GPUs, and even smartphones.

This March, QVAC released a LoRA fine-tuning framework for Microsoft's BitNet 1-bit model; in April, they launched the SDK; and in June, they made an open-source implementation of Google's research institute's memory compression algorithm TurboQuant. There is also a local AI desktop application called Workbench.

A stablecoin company is doing edge inference optimization, which sounds absurd, but these things are indeed on GitHub.

Tether's money also took a more roundabout route. In December 2024, it made a strategic investment of $775 million in the video platform Rumble; in June this year, Rumble completed the acquisition of the German publicly listed company Northern Data, obtaining an 85.2% stake.

Northern Data possesses over 22,000 GPUs in its European cloud business, plus about 250 MW of data center capacity expected by 2027. It raised its revenue guidance for 2026 from €130 million to €150 million up to €170 million to €190 million, signing a multi-year GPU cloud contract worth $270 million with Together AI as the client.

A profit made from stablecoins, bypassing a video company, ultimately transformed into a GPU data center in Europe.

Crypto.com’s $70 million was also not just for holding. Marszalek began building a team after acquiring AI.com in April 2025. The product was launched post-Super Bowl, offering personal AI agents capable of sending messages on behalf of users, executing operations across applications, and trading stocks. He simultaneously serves as CEO of two companies.

On the money side, Paradigm closed its fourth fund of $1.2 billion on July 8, clearly specifying AI and robotics. This firm, managing nearly $12 billion in crypto venture capital, has already invested in drone delivery company Zipline and space defense company True Anomaly.

The most ironic aspect is OpenRouter.

The founder Alex Atallah is a co-founder and CTO of OpenSea. He left in early 2023 to create something that acts as a switch for over 400 large models: developers make one API call, and it decides whether that request lands on GPT, Claude, or a certain open-source model.

This May, OpenRouter raised $113 million in Series B funding, achieving a valuation of $1.3 billion. It has 8 million registered developers, processing 250 trillion Tokens weekly.

An industry that spent a decade proving "no need for intermediaries" produced its most successful company as an intermediary for models.

On the other hand, the money in crypto is visibly being drawn out. In April 2026, global crypto venture capital only invested $659 million in a month, the lowest in two years, down 75% year-on-year.

Why is the money leaving? Paradigm hasn't said much about it, but the books on the LP side are clear. After the bursts of FTX, Terra, and Three Arrows, institutional trust in crypto funds has not been restored. That batch of projects that received high valuations in the last cycle, produced neither income nor accumulated real users.

In AI, at least there are bills to be verified.

The last time the industry collectively changed direction was after the ICO crash in 2018, when everyone turned to DeFi.

This time it was like changing rooms in their own home.

This time, they moved out.

03 But the report cards are polarized

Not everyone transformed successfully.

Canaan is a counterexample. This company, which produces Bitcoin mining machines, also ventured into AI chips, contributing about $900,000 in revenue to that sector in 2024. The total revenue for the company that year was $269.3 million.

$900,000 against $269.3 million. And this department consumed 15% of the entire company's operating expenses.

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

The group of mining companies is also stratifying. Those with electricity, power grid access, and ready-made data centers secured long-term contracts with Microsoft and Google; others with only mining machines and a press release see little follow-up after the announcement.

The same MW, in the hands of a company with delivery capacity, is a fifteen-year lease; in the hands of a company without, it’s just a slide.

The distinguishing criterion is actually one: is anyone really paying for it?

Galaxy has CoreWeave's rent, Crusoe has Microsoft's and OpenAI's campuses, OpenRouter has 8 million developers tuning its API. Canaan has a $900,000 annual revenue, then it calculated that account clearly itself.

04 Summary

This batch of companies are not "crypto's AI"; they are just AI.

What truly moved over is not the technology. It is a group that knows how to raise billions when no one believes in you, and how to electrify a barren land in just a few months. These skills have little to do with blockchain, but coincidentally were honed in the crypto industry.

It also needs to be made clear that walking out does not equal winning. Crusoe's Abilene Microsoft campus won't be electrified until mid-2027, Crypto.com's AI product has just launched, and Rumble, after acquiring Northern Data, needs to first prove its operational capabilities.

None of these are up for acceptance yet. One thing is certain: the bets have been placed, and they are on someone else's table.

Ten years ago, they said they would eliminate intermediaries.

Now they are acting as intermediaries for AI.