AMD has officially entered a multi-year strategic partnership to supply OpenAI with 6 gigawatts of compute capacity. This massive infrastructure deal, valued at tens of billions of dollars, positions AMD as a critical player in the global AI hardware race and significantly bolsters its long-term revenue outlook.
Powering the Future of AI Infrastructure
The agreement centers on the deployment of multiple generations of AMD’s Instinct GPUs, starting with the upcoming Instinct MI450. The 6-gigawatt capacity is enough to power approximately 4.5 million homes. According to the rollout schedule, OpenAI will receive the initial gigawatt of capacity in the second half of 2026, coinciding with the deployment of the MI450 architecture.
AMD anticipates that the MI450 series will outperform Nvidia’s Rubin CPX offerings. These performance gains are expected to be driven by hardware and software refinements developed in direct collaboration with OpenAI. Currently, OpenAI already utilizes AMD’s MI355X and MI300X GPUs for large language model inference, citing their superior memory capacity and bandwidth as key advantages.
Strategic Equity Stakes and Stock Milestones
Beyond hardware supply, the contract includes a significant financial incentive structure. AMD has granted OpenAI the option to acquire up to 160 million shares, representing a 10% stake in the company. Vesting for these shares is tied to specific performance milestones:
- The first tranche vests upon the initial 1-gigawatt deployment.
- Subsequent tranches vest as OpenAI scales toward the full 6-gigawatt capacity.
- Final equity vesting is contingent on AMD’s stock reaching $600 per share.
The market reaction to the announcement was immediate, with AMD shares surging nearly 35% following the disclosure.
OpenAI’s Aggressive Expansion Strategy
This partnership is a cornerstone of OpenAI’s broader plan to secure the compute resources necessary for its “Stargate” data center initiative, which aims for a total capacity of 7 gigawatts across five new facilities.
Dr. Lisa Su, CEO of AMD, emphasized that the collaboration aims to facilitate the world’s most ambitious AI buildout. OpenAI CEO Sam Altman echoed this sentiment, describing the deal as a vital step in unlocking the full potential of AI.
The Compute Arms Race
OpenAI’s deal with AMD is part of an intensive effort to diversify its supply chain. Recent weeks have seen a flurry of activity from the AI giant, including:
- A $100 billion investment agreement with Nvidia for at least 10 gigawatts of capacity.
- A $10 billion collaboration with Broadcom for custom AI chip development.
- Expansion of the Stargate initiative alongside Oracle and SoftBank.
- New supply agreements with Samsung Electronics and SK Hynix for specialized DRAM memory.
As OpenAI races to build out its infrastructure, this multi-front approach signals a shift toward a more modular and competitive AI hardware ecosystem, moving beyond reliance on a single vendor.