When Core Scientific signed a $3.5 billion deal to host an AI (AI) data center earlier this year, it wasn’t chasing the next Crypto token. Once known for its huge fleet of Bitcoin mining rigs, the company is now part of a growing trend: converting energy-intensive mining operations into high-performance AI facilities.
Bitcoin miners like Core, Hut 8 (Hut) and Terawulf (Wulf) are replacing ASIC machines (dedicated Bitcoin mining computers) in GPU clusters driven by the temptation of AI’s explosive growth and the harsh economy of crypto mining.
Power play
It is no secret that Bitcoin mining requires a wide range of energy. This is the biggest cost of minting new digital assets.
Back in the 2021 Bull Run, when the hashrate and difficulties on the Bitcoin network were low, miners were making like bandits with a 90% margin. Then came the brutal code winter and half events, cutting mining fees in half. In 2025, miners are struggling to survive on the thin razor rim, as hashrate and energy prices skyrocket.
However, the biggest input cost, the need for power, is a disguise blessing for these miners who needed different strategies to diversify revenue streams.
With competition for the mining intensified, miners continued to source more machines to stay on the water, which required more electricity and more electricity at a cheaper price. Miners have invested heavily in securing these low-cost energy sources, such as hydroelectric power generation and chained natural gas sites, and have developed expertise in managing high-density cooling and electrical systems that were polished during the crypto boom of the early 2020s.
This is what caught the attention of AI and cloud computing companies. Bitcoin relies on professional ASICs, but AI thrives on versatile GPUs like NVIDIA’s H100 series. This requires a similar high power environment, but is necessary for parallel processing tasks in machine learning. Instead of building a data center from scratch, taking over a mining infrastructure that already has power has become a faster way to increase the speedup of AI-related infrastructure.
Essentially, these miners are retrofitting, not just pivots.
The cooling systems built during Crypto Boom, low-cost energy contracts, and high-power density infrastructures currently serve the new objectives that are being fed into AI models for companies such as Openai and Google.
Companies like Crusoe Energy sell mining assets to focus solely on AI and deploy GPU clusters in remote, energy-rich locations that reflect the decentralized ethos of crypto, but are now focusing on AI hyperscalar.
Terraforming AI
Bitcoin mining effectively “terraforms” the terrain for AI calculations by building the scalable, power-efficient infrastructure that AI desperately needs.
As Nicholas Gregory, director of Fragrant Prosperity, put it, “You can argue that Bitcoin paved the way for digital dollar payments, as you can see in USDT/Tether. It also looks like a Bitcoin Terra Formation data center for AI/GPU computing.”
This existing “terraforming” allows miners to rapidly renovate their facilities in less than a year, often compared to the multi-year timelines of traditional data center builds. Companies like Crusoe Energy sell mining assets to focus solely on AI and deploy GPU clusters in remote, energy-rich locations that reflect the decentralized ethos of crypto, but are now focusing on AI hyperscalar.
Higher return
In reality, this means that miners can turn their facilities upside down within a year. This is faster than the multi-year timeline for the new data center.
However, AI is not a cheap upgrade.
Bitcoin mining setups are relatively modest, with costs ranging from $300,000 to $800,000 per megawatt (MW) excluding ASICs, allowing for rapid scalability depending on the market cycle. Meanwhile, AI infrastructures have significantly higher CAPEX due to the need for advanced liquid cooling, redundant power systems, and the GPU itself. Despite the sudden upfront costs, AI provides miners with up to 25 times more revenue per kilowatt-hour than Bitcoin mining, making it economically attractive amid rising energy prices and declining cryptocurrency.
A niche industry worth billions
As AI continues to surge and crypto profits tighten, Bitcoin mining can become a niche game. Particularly, the next task in 2028 can make many tasks unprofitable without breakthroughs in efficiency or energy costs.
The forecast shows that the global crypto mining market has grown to $3.3 billion by 2030, but at a 6.9% CAGR, billions will be hidden by the exponential expansion of AI. According to a KBV study, the global AI in the mining market is projected to reach $43.594 billion by 2032, expanding at a combined annual growth rate (CAGR) of 40.6%.
As investors are already seeing dollar signs in this shift, the broader trend suggests that the future is a complete conversion to hybrid or AI. A stable contract with Hypescalar promises a longer life than the crypto boom bust cycle.
This evolution not only reuses idle assets, but also highlights how yesterday’s crypto frontier is forgering tomorrow’s AI empire.
