
Key Points
AMD and Core Scientific sign AI infrastructure pact
Core Scientific and Advanced Micro Devices (AMD) have entered a partnership aimed at expanding AI-focused data-center capacity in the United States. The agreement gives AMD access to more than 500 megawatts of data-center infrastructure beginning in 2027. The deal also includes the option to scale that capacity up to 2.5 gigawatts, positioning both companies to support large AI compute demands.
The partnership is structured to secure long-term power and facility resources for AMD’s AI solutions. By reserving significant capacity in advance, the companies are targeting the needs of end customers that require high-performance computing environments for AI workloads. The arrangement highlights the growing importance of dedicated infrastructure planning in the AI supply chain.
Scope of data-center capacity and timing
Under the agreement, more than 500 megawatts of U.S. data-center capacity are expected to come online for AMD’s use starting in 2027. This capacity is intended to be AI-ready, supporting high-density compute deployments. The contract includes rights to expand the total capacity commitment up to 2.5 gigawatts over time, subject to future scaling decisions.
The focus on U.S. infrastructure underscores the priority placed on domestic data-center development for AI applications. Beginning in 2027, the capacity is intended to serve end customers that deploy AMD-based AI systems, adding to the available footprint for large-scale model training and inference.
Technology collaboration and deployment plans
Beyond reserving capacity, AMD and Core Scientific plan to work together on the physical design of the data-center infrastructure. The collaboration covers how facilities will be built and configured to efficiently host advanced AI hardware. This includes considerations such as power delivery, cooling, and rack layouts suitable for high-performance chips.
The deployment roadmap spans AMD Instinct GPUs, EPYC CPUs, and the ROCm software stack. These components form the core of AMD’s AI and compute platform, combining accelerators, general-purpose processors, and an open software environment. By integrating these technologies into dedicated facilities, the partners aim to provide an optimized environment for AI deployments.
Rationale amid accelerating AI deployments
AMD framed the agreement as a direct response to rapidly accelerating AI deployments. Mathew Hein, senior vice president and chief strategy officer of corporate development at AMD, said that bringing the required compute online demands trusted infrastructure partners with sufficient scale and power to support the next era of AI.
The partnership with Core Scientific is designed to address that need by tying together large-scale power capacity, specialized data-center design, and AMD’s AI hardware and software. It reflects an effort to align chip supply with purpose-built infrastructure so that customers can deploy advanced AI systems as demand grows from 2027 onwards.
Key Takeaways
- 01The deal secures a sizable block of future U.S. data-center capacity specifically aligned to AMD’s AI roadmap.
- 02Collaboration on facility design indicates that AI infrastructure needs are shaping how new data centers are engineered.
- 03The option to scale from 500 megawatts to 2.5 gigawatts highlights expectations for substantial growth in AI compute demand.
References
- https://cnbc.com/2026/07/27/crypto-stocks-rally-thanks-to-rotation-from-ai-infrastructure-bitcoin-miners-lag.html
- https://www.businesswire.com/news/home/20260727256371/en/Core-Scientific-and-AMD-Announce-Infrastructure-Partnership
- https://stocktwits.com/news-articles/markets/equity/corz-stock-rebounds-after-amd-partnership-unlocks-multi-gigawatt-ai-expansion/cZZCNLnRJGv
- https://www.tradingview.com/news/stocktwits:00c876c0c094b:0-corz-stock-rebounds-after-amd-partnership-unlocks-multi-gigawatt-ai-expansion/