Bitcoin Miners Fuel Surge In AI Data Centre Expansion - Arabian Post

Bitcoin mining operations are undergoing a transformation as firms leverage their substantial electricity capacity and infrastructure to meet booming demand for artificial intelligence compute. Companies that once focused solely on mining are now converting parts of their facilities into AI data centres, pressing for more stable revenue amid tightening profit margins in crypto.

CoreWeave's $9 billion acquisition of Core Scientific stands out as a major move in this trend, enabling CoreWeave to control more high-performance data centre capacity and eliminate costly lease obligations. The deal is projected to yield cost savings of around $500 million annually by 2027. Both companies, originally rooted in cryptocurrency mining, have shifted significantly towards hosting AI workloads.
Several other miners are pursuing similar strategies. Bitfarms, based in Toronto, is evaluating multiple sites in North America for conversion to AI and high-performance computing use, hiring consulting firms to assist with computing and AI strategy. Its CEO has said that while their bitcoin mining operations will continue, long-term contracts with AI customers offer steadier cash flows.

Mining firms are well placed for this pivot because they already maintain power-dense infrastructure, cooling systems, and land suited for large electrical loads. These traits, developed to support energy-intensive mining rigs, are in high demand among enterprises seeking to train and deploy large-scale AI models. As energy costs rise and regulatory pressure grows, the economics of GPU-based AI compute are becoming more attractive compared with proof-of-work crypto mining.

Core Scientific, after emerging from bankruptcy, now channels most of its power capacity-roughly 800 megawatts-into data-centre computing deals, with the remainder still used for bitcoin mining. It has secured long-term leases for its data centre space and aims to upgrade existing warehouses with advanced cooling and increased power density. Such upgrades are essential, as AI workloads require more reliability, better thermal management and higher electrical capacity than many mining operations.

Investors are responding. Firms like Iren are expanding their GPU inventory and positioning themselves to offer AI cloud services at scale. These moves promise higher revenue per kilowatt hour compared to traditional mining, even after factoring in the higher upfront costs.

Challenges remain. AI data centres require robust network connectivity, stricter reliability, and often far more complex cooling solutions compared to bitcoin mining. Some mining facilities are located in regions without the fibre, infrastructure or regulatory environment needed for hyperscale AI deployment. Even for those that have suitable electricity and cooling, meeting customer standards for uptime, security, and latency demands significant upgrades.

Energy supply and sustainability also pose concerns. As more miners turn to AI, securing low-cost, clean power becomes crucial. Some operators are exploring renewable sources, off-grid energy, or sourcing idle capacity. For example, firms are using surplus energy during off-peak periods or tapping into stranded natural gas sites. Such strategies not only improve margins but also help address public and regulatory pressure over environmental impact.

Grid stability is emerging as an unexpected benefit of the crossover. Because AI workloads often have variable demand, and bitcoin mining can be dialled up or down relatively quickly, miners can act as flexible consumers. By switching load between AI compute and mining as conditions dictate, facilities can help absorb demand spikes or make use of surplus renewables, easing stress on the power grid.

MENAFN14092025000152002308ID1110059489