Core Scientific will allocate up to 2.5 GW of data centers to AMD for AI infrastructure.
A major player in the mining sector, Core Scientific, has entered into a strategic agreement with the American processor and graphics chip manufacturer AMD. Under the terms of the partnership, the mining company commits to providing AMD with up to 2.5 gigawatts of computing capacity from its data centers. These resources will be directed toward deploying client artificial intelligence systems.
This deal marks an important stage in the transformation of infrastructure originally built for cryptocurrency mining into high-tech data centers for AI workloads. Core Scientific, known for its large-scale mining farms, now acts as a supplier of critically important capacity for one of the leaders in the semiconductor industry. The volume of 2.5 GW is a colossal figure, comparable to the energy consumption of an entire small city, underscoring the ambition of AMD's plans to expand its presence in the artificial intelligence segment.
From a market perspective, this agreement is a logical continuation of the trend of convergence between mining and AI. Companies that own large data centers and have experience managing energy-intensive systems are becoming key partners for technology giants. Core Scientific is not just diversifying its business but also creating a stable source of income that is less dependent on cryptocurrency volatility. For AMD, this is an opportunity to quickly scale up access to necessary infrastructure without the capital expenditure of building its own facilities.
Expert Commentary: From my point of view, this agreement is a powerful signal for the entire market. It proves that infrastructure for Bitcoin mining holds immense value beyond the crypto ecosystem. Investors who were previously skeptical of mining companies due to their dependence on BTC price now see a new monetization model. Core Scientific is successfully reshaping its business, becoming the "backend" for the AI revolution. In the coming years, we will likely see even more such deals, as demand for computing power for training and inference of neural networks will only continue to grow.