Several companies this year that once focused on Bitcoin mining are now moving into artificial intelligence (AI) infrastructure. This change is part of a larger trend. The demand for AI computing power has grown faster than firms can build it. At the same time, profits from traditional crypto mining have fallen due to lower rewards and rising costs.

Mining companies are now using their energy infrastructure and data centers for AI workloads. This shift goes beyond just mining Bitcoin.

One clear example of this shift is Hut 8 Corp., a company known for Bitcoin mining. The company made a major announcement. It signed a 15-year lease agreement with cloud infrastructure provider Fluidstack worth about $7 billion to provide data center space and power for AI computing. This deal marks a big step in the company’s strategy to pivot from crypto mining to AI infrastructure.

Hut 8’s Strategic AI Deal

Hut 8’s $7 billion agreement involves its River Bend campus in Louisiana. The lease covers 245 megawatts (MW) of computing capacity that will be used for AI workloads. This capacity will support high-performance computing clusters, which are needed to run advanced AI models. The first data hall is scheduled to be operational by early 2027.

The lease is structured for 15 years, giving Hut 8 a long-term revenue stream from AI infrastructure. The company also has options to expand capacity.

Under the agreement, the site could grow far beyond the current 245 MW. FluidStack has the right of first offer for up to an extra 1,000 MW at River Bend if expansion occurs.

This deal is supported financially by Google, which acts as a financial backstop. A financial backstop means Google guarantees lease payments if the tenant cannot pay them. This reduces the financial risk for Hut 8 and makes the project more bankable. Major banks like J.P. Morgan and Goldman Sachs are also involved in financing.

Fluidstack will operate the data center, and the AI models that run there are expected to support work by Anthropic, a major AI research company. This partnership between Hut 8, Fluidstack, and Anthropic shows how crypto mining firms are now engaging with the AI ecosystem.

Asher Genoot, CEO of Hut 8, noted:

“Scaling frontier AI infrastructure is, at its core, a power challenge. Hut 8’s power-first, innovation-driven development model enables us to originate and develop greenfield data center sites at the pace and scale required by leading model developers.”

Why Crypto Miners Are Pivoting

Traditional Bitcoin mining uses specialized machines called ASICs. These machines solve complex math problems to earn new Bitcoin.

Over time, rewards from mining have become smaller. For example, after the 2024 Bitcoin halving event, block rewards were cut by half. This made mining less profitable for many companies.

Mining also uses a lot of energy. It consumes ~173 TWh annually (2025), or ~0.5% global electricity, ~10 GW continuous power equivalent to Poland. Bitcoin mining, for instance, has been using more power each year since 2015, from about 0.15 TWh to over 170 TWh in 2025.

Moreover, crypto miners often build large data halls with high-capacity, 100+ MW power lines, cooling systems, and backup generators. These systems are expensive to build, but they are well-suited for another purpose: hosting AI computing workloads. High-performance AI computing also needs a lot of power, cooling, and space for hardware.

AI computing workloads rely heavily on graphics processing units (GPUs). GPUs are different from ASICs but require similar infrastructure. They need high power delivery and advanced cooling. Crypto miners already own these facilities. Many are now turning them into AI data centers to avoid letting them sit idle or lose money.

This strategy also helps companies diversify revenue. Mining revenue can be volatile. It depends on the price of Bitcoin, mining difficulty, and energy costs.

AI infrastructure leasing, on the other hand, can offer predictable long-term income. A 15-year lease, like Hut 8’s deal, can provide stable revenue even if crypto markets fluctuate.

From Hash Rate to Compute Power: A Sector-Wide Shift

Hut 8 is not alone in this transformation. Other former crypto mining firms are pursuing similar paths. For example, CoreWeave, another company with roots in crypto mining, shifted its operations to focus on AI compute services.

By mid-2025, CoreWeave reported revenue of more than $1.2 billion and a valuation of about $48 billion. The firm now runs dozens of data centers with GPU capacity for AI workloads.

In addition, companies like TeraWulf have partnered with Fluidstack in long-term agreements to provide hundreds of megawatts of compute power in New York. Such deals reflect growing interest in AI and high-performance computing (HPC) infrastructure.

The shift from crypto mining to AI computing is also fueled by wider market expectations. Industry analysts project strong growth in AI infrastructure demand.

Some forecasts suggest the global AI infrastructure market may grow over 20% each year until the early 2030s. This growth is fueled by demand for generative AI and other advanced models. This growth requires substantial investment in data centers, high-power computing hardware, and supporting networks.

One report says the AI infrastructure market could grow from about $47 billion in 2024 to over $499 billion by 2034. However, these numbers differ depending on the source and method used. Such expansion suggests that large computing capacity will be needed to support model training, inference, and cloud-based AI services.

How Wall Street Is Pricing the AI Pivot

Hut 8’s announcement had a noticeable impact on its stock price. After news of the $7 billion lease broke, the company’s shares climbed significantly in pre-market trading.

In one report, shares surged about 21% following the announcement. Score gains across the year, 79%, also reflect changing investor sentiment toward the company’s growth strategy.

The AI infrastructure strategy could help stabilize Hut 8’s financial performance. Analysts note that long-term leasing deals like this can produce predictable net operating income.

Hut 8 might earn about $6.9 billion in net income during the first 15 years of the lease. If they exercise renewal options, this could rise to nearly $17.7 billion.

Stable companies offering long-term data center services can secure financing more easily. They also tend to keep investment-grade credit ratings. This is different from the ups and downs of crypto mining revenue. Earnings rely a lot on Bitcoin prices and network difficulty.

High Stakes, High Power, High Competition

Despite the growing interest, the shift from crypto mining to AI infrastructure is not without challenges. Building and operating AI data centers requires expertise in data center management, network architecture, and cloud services. Some crypto mining firms need to invest in new equipment, such as GPUs and networking gear, to support AI workloads.

Competition in this sector is also intense. Big cloud providers and large data center companies, like Amazon, Microsoft, and Google, have a lot of experience and money for AI computing. Crypto-turned-AI firms must find a way to compete effectively by offering differentiated services or cost advantages.

There are also questions about energy use and sustainability. High-performance AI computing requires substantial electricity and cooling. Companies must balance growth in capacity with environmental and regulatory concerns.

As global demand for data center power increases, firms may face pressure to use renewable energy or adopt energy-efficient designs.

The Rise of Hybrid Crypto–AI Business Models

The trend of crypto miners pivoting to AI infrastructure shows how industries adapt to changing economic and technology landscapes. Firms like Hut 8 are using their existing strengths—power infrastructure and data centers—to meet rising demand for AI computing. Deals like the $7 billion, 15-year lease with Fluidstack and support from Google highlight this shift.

Some analysts believe that combining crypto mining expertise with AI infrastructure services could create new hybrid business models. These might include GPU-as-a-Service offerings, HPC leasing, and partnerships with major AI developers. For instance, Hut 8’s GPU-as-a-Service subsidiary has already deployed more than 1,000 Nvidia H100 GPUs to serve AI clients.

Other companies are making similar moves as the market for AI capacity expands rapidly. While challenges remain, the pivot could help stabilize revenue and position former miners in a growing and increasingly essential part of the technology economy.