AMD and OpenAI formed a strategic partnership to deploy 6 gigawatts of graphics processing units for artificial intelligence infrastructure beginning in the second half of 2026. The collaboration starts with 1 gigawatt of next-generation MI450 accelerators and expands across multiple Instinct GPU generations. AMD issued OpenAI a warrant for up to 160 million shares of common stock that vests as deployment milestones are achieved and share-price targets are met.
The MI450 accelerators deliver 40 petaflops of compute performance with 432 gigabytes of HBM4 memory and 19.6 terabytes per second bandwidth. OpenAI will work with AMD as a core strategic compute partner to optimize product roadmaps and expand hardware collaboration beyond the MI300X and MI350X series. AMD Chief Financial Officer Jean Hu stated the agreement should generate tens of billions of dollars in revenue while accelerating infrastructure development.
The partnership combines AMD leadership in high-performance computing systems with OpenAI research advancements in generative artificial intelligence. OpenAI Chief Executive Officer Sam Altman said the collaboration builds the compute capacity needed to realize the technology's full potential. The agreement aligns strategic interests between both companies to advance the artificial intelligence ecosystem and meet growing global demands for computational infrastructure.
The MI450 accelerators deliver 40 petaflops of compute performance with 432 gigabytes of HBM4 memory and 19.6 terabytes per second bandwidth. OpenAI will work with AMD as a core strategic compute partner to optimize product roadmaps and expand hardware collaboration beyond the MI300X and MI350X series. AMD Chief Financial Officer Jean Hu stated the agreement should generate tens of billions of dollars in revenue while accelerating infrastructure development.
The partnership combines AMD leadership in high-performance computing systems with OpenAI research advancements in generative artificial intelligence. OpenAI Chief Executive Officer Sam Altman said the collaboration builds the compute capacity needed to realize the technology's full potential. The agreement aligns strategic interests between both companies to advance the artificial intelligence ecosystem and meet growing global demands for computational infrastructure.
