AMD and OpenAI Announce Strategic Partnership for 6 Gigawatt GPU for Generative AI Infrastructure

AMD and OpenAI announce a strategic agreement totaling 6 gigawatts of AMD Instinct GPUs to support OpenAI’s next-generation AI infrastructure. The initial phase of the agreement includes the deployment of 1 gigawatt of AMD Instinct MI450 GPUs planned to begin in the second half of 2026, with continued supply to the next generation of GPUs gradually.

Agreement Details

In this agreement, OpenAI appoints AMD as its primary strategic computing partner for deploying rack-scale AI solutions using multiple generations of Instinct GPUs. The technical collaboration will focus on optimizing AMD’s product roadmap and multi-generation hardware–software integration that has already started with the MI300X and MI350X series, and will continue to MI450 and future generations.

As part of long-term alignment, AMD grants options (warrants) for up to 160 million common shares to OpenAI. The granting of these options will be phased based on milestone achievements—starting with the initial 1 GW deployment—and dependent on a combination of AMD stock price targets and OpenAI’s technical and commercial milestones that enable larger scale deployments up to 6 GW.

Dr. Lisa Su, Chair and CEO of AMD, expressed enthusiasm about the partnership, emphasizing that this collaboration combines AMD’s computing excellence and OpenAI’s AI research to enable the building of very large-scale AI.

Sam Altman, CEO of OpenAI, highlighted that AMD’s computing capacity support will accelerate OpenAI’s ability to realize the full potential of generative AI technology and bring benefits faster to a broad audience. Greg Brockman, President of OpenAI, added that AI progress requires deep collaboration across all technology layers, and working with AMD enables scaling and broader deployment of AI tools.

Jean Hu, EVP and CFO of AMD, estimated that this partnership could generate tens of billions of dollars in revenue for AMD while accelerating the development of OpenAI’s computing infrastructure and increasing shareholder value through impact on AMD’s non-GAAP earnings per share.

This agreement confirms the industry trend where chip providers and AI companies form strategic partnerships to meet the ever-growing demand for computing. For OpenAI, access to large-scale capacity from AMD enables training and deployment of larger and faster models. For AMD, multi-generation contracts like this strengthen their position in the HPC/AI market and open long-term revenue opportunities.

The initial 1 GW phase in the second half of 2026 is the first practical milestone; further developments will depend on achieving technical, commercial milestones, and market conditions.

The AMD–OpenAI collaboration marks a major step in building large-scale AI infrastructure: combining AMD’s chip design and system integration capabilities with OpenAI’s intensive computing needs. If successful, this agreement will not only accelerate ambitions for generative AI products and services but also drive the global AI ecosystem toward higher capacity and performance.