The hundred-billion-dollar race to build artificial-intelligence data centers as fast as possible has found a new target: bitcoin miners.
Across the country, so-called “mega-scalers” looking to build data centers in the gigawatt range have unlocked hundreds of billions of dollars to support an unprecedented effort to increase both the intelligence of AI products and deploy them widely. That means companies such as Microsoft, Google, Oracle, OpenAI and Meta are pouring money into projects across the country.
But the bottlenecks are growing, with mounting local resistance to projects, concerns over water and electricity usage, and yearlong delays to connect to the wider grid for the massive power needed to operate.
These companies, designed with the computer power needed to produce new bitcoin, already have a lot of what the hyper-scalers need: Massive, secured electrical power capacity in areas with low rates, industrial-level cooling systems, large campuses and high-bandwidth networking potential.
That’s not to say it’s an easy conversion, as bitcoin mines use different technology than AI data centers. AI data centers also need multiple levels of redundancy and often denser cooling and power requirements. The transformation can cost tens of millions of dollars or more, experts say, but it offers the most attractive quality for data center developers — speed.
“Standing up new substations and transmission lines can take years, and interconnection queues are only getting longer,” said Wayne Highfield, director of operations at Megawatt, a large-scale mining company based in Indiana. “Partnering with existing mining operators can shorten timelines, put available energy to work immediately, and create more pathways to get projects online.”
AI shift fueled by demand
The desire for speed has already begun playing out across the country.
Bitcoin miner Bitfarms Ltd. announced in November that it was converting its 18-megawatt bitcoin mining facility in Washington state to support AI workloads after signing a $128 million agreement with what it called a “large, publicly traded” provider of data centers.
Ben Gagnon, CEO of Bitfarms, said at the time that there were compelling reasons for the company to pursue a “GPU-as-a-service” model and ultimately forgo mining.
“Despite being less than 1% of our total developable portfolio, we believe that the conversion of just our Washington site to GPU-as-a-Service could potentially produce more net operating income than we have ever generated with Bitcoin mining,” he said, adding it could provide the company with a strong cashflow foundation.
Bitfarms is not the only miner to cash in on the AI tsunami.
CleanSpark Inc., which bills itself as “America’s bitcoin miner,” appointed a senior executive in charge of data centers in October as part of its effort to expand beyond bitcoin.
Former or current bitcoin miners such as TeraWulf, Soluna Holdings, Hut 8 and others have inked deals over the last year to convert existing facilities, build new AI data centers on existing sites or co-locate AI with mining operations.
Investment firm VanEck, which also offers bitcoin and related investments, said in a January report that mining activity has dropped by about 2% and the overall “hash rate” dropped 6%, a decline it attributed to miners powering down rigs to instead service demand for AI data centers.
“We expect AI data center demand to persist over the coming years, with a (+24%) CAGR through 2030, and expect bitcoin miners to increasingly devote power resources to servicing the buildout of artificial intelligence,” the company said in a blog post.
Digital asset investment manager CoinShares said the pivot to AI data centers is attractive because it offers more stable, predictable revenue — and at a much higher margin. How much? About three times the return of bitcoin on a per-megawatt basis.
By October, bitcoin firms announced about $65 billion worth of contracts with AI firms or data center operators, CoinShares reported. Of the six publicly-traded companies that have announced AI projects so far, CoinShares anticipates that bitcoin mining revenues will decline from about 85% in 2025 to under 20% by the end of 2026.
Morgan Stanley said in a September report that bitcoin mining companies have about 6.3 gigawatts of operational, large sites and another 2.5 gigawatts under construction, making them the “fastest way to obtain electricity with the lowest execution risk” for AI companies. Morgan Stanley said if mining sites were transformed into data centers, it would create about $5 to $8 of equity value per watt, far more than bitcoin.
Private consulting firm WiseGuy Reports expects to see more of the bitcoin-to-AI pivot — especially among the largest, publicly-traded firms — as more value is being placed on the number of megawatts these firms control, opposed to the power they can bring to bitcoin mining operations.
“As of early 2026, the trend has moved beyond experimentation into a full-scale land grab for power,” the report said. “Hyperscalers are so desperate for capacity that they are signing multi-billion-dollar pre-lease agreements before the retrofits are even complete.”
Another pressure on bitcoin miners is the price of bitcoin, which has steadily dropped from an October high of around $124,000 per bitcoin to around $70,000 as of publication time — putting it below the cost to mine it.
Data center conversion potential
How many potential sites across the country are capable of being converted to AI? There is no exact number.
The Energy Information Administration estimated about 137 bitcoin mines were operational across the country in 2023 — about 2.3% of the total American energy consumption. However, not every site will be suited to convert to AI demand, Wise Guy Reports said, and not every bitcoin miner will have the resources to convert their operations over to AI operations.
But for the companies that show they have the land, power and resources, there will be no shortage of interested parties willing to fund the buildout.
“This ‘build-to-suit’ demand creates a ‘de-risked’ environment for miners; if they can prove they have the power and the ability to build a Tier 3 shell, the tenants are already waiting at the door with open checkbooks,” the report said.
The rapid rise of generative AI platforms such as OpenAI’s ChatGPT, Anthropic’s Claude, Google’s Gemini and others has meant hundreds of millions of users per day are generating text and images. When it comes to widespread adoption, though, cracks have started to appear.
Small businesses collectively are not racing to adopt AI tools, with overall adoption having slowed in some cases and having stopped in others, according to a fall 2025 Census Bureau Business Trends and Outlook Survey, which is conducted once every two weeks surveying 200,000 small businesses. The percentage of small-business owners who report using AI to produce goods or services has grown since the survey launched in 2023, from an initial 3.7% to 9.7% in September, according to the survey, first reported by the Apollo Academy. The companies that are adopting the technology vary substantially by size, though.
Source: “Data center developers are zeroing in on a new target“
