OpenAI has entered into a contract with AMD for GPUs to be utilized in its next-generation AI infrastructure in a deal valued at tens of billions of dollars, granting the ChatGPT developer the option to acquire a 10 percent stake in its supplier as part of the agreement.
As part of the agreement, OpenAI will implement six gigawatts of AMD GPUs within its infrastructure gradually through various generations of the chip manufacturer’s silicon. The initial shipment of AMD’s upcoming Instinct MI450 GPUs is scheduled for delivery in the second half of 2026.
In a joint announcement, both companies emphasized that they would also exchange technical knowledge to “enhance” product roadmaps, mentioning that the partnership was based on existing collaborations. They described it as one that would facilitate massive AI implementations and promote the complete ecosystem.
In conjunction with the agreement, AMD granted OpenAI a performance-based warrant that allows it to purchase up to 160 million AMD shares in increments at a predetermined price. This activates as milestones are reached and may ultimately total approximately a 10 percent share.
AMD’s chair and CEO, Lisa Su, mentioned that the “collaboration combines the strengths of AMD and OpenAI to foster a genuine win-win, facilitating the most ambitious AI development and enhancing the overall AI ecosystem.”
CFO Jean Hu estimated that it would “generate tens of billions of dollars in revenue for AMD, while speeding up the development of OpenAI’s AI infrastructure.”
OpenAI co-founder and CEO Sam Altman emphasized that the agreement is “a significant advancement in establishing the computational power required to unlock AI’s complete potential,” further stating, “AMD’s dominance in high-performance processors will allow us to speed up development and deliver the advantages of cutting-edge AI to all more quickly.”
A Major Victory for AMD
Jack Gold, president and principal analyst at J. Gold Associates, characterized the agreement as a “significant victory for AMD,” noting that it serves as “an affirmation that AMD has achieved competitiveness with Nvidia’s processing capabilities.”
He observed that it reflects OpenAI’s acknowledgment of the necessity to diversify its processor sources while it continues to grow its data centers. The latest Nvidia GPUs are on allocation, as production is exceeding available supplies. By securing AMD chip supplies through this dedication and investment, OpenAI can maintain its extensive expansion efforts.
The analyst urged other “major AI companies” to “emulate this and implement AMD-powered data centers.”
