AMD has signed a multi-year chip supply agreement with OpenAI, a deal projected to generate tens of billions in revenue for the chipmaker and accelerate its momentum in the artificial intelligence industry. Under the terms of the agreement, AMD will supply six gigawatts of compute capacity to OpenAI. This amount of power is enough to supply up to four and a half million homes. The supply will span multiple generations of AMD’s Instinct GPUs, commencing with the new Instinct MI450 series.
OpenAI is scheduled to receive the first gigawatt of this capacity in the second half of 2026, which aligns with the planned deployment of the new MI450 chip. AMD asserts that its MI450 series will outperform comparable offerings from Nvidia, specifically the Rubin CPX, through a combination of hardware and software enhancements. Many of these improvements will be developed with direct input from OpenAI. AMD’s current generation GPUs, the MI355X and MI300X series, are already utilized by OpenAI for some workloads and are recognized as strong performers for AI inference in large language models due to their substantial memory capacity and bandwidth.
A significant component of the deal involves AMD granting OpenAI an option to purchase up to 160 million shares of AMD stock, equating to a ten percent stake in the company. The first portion of these shares will vest upon the initial one gigawatt deployment. Additional portions will vest as OpenAI purchases compute capacity up to the full six gigawatts. The vesting of OpenAI’s stake is also directly linked to increasing AMD’s stock price, with the final tranche vesting only when the share price reaches six hundred dollars. For context, AMD shares closed at one hundred sixty-four dollars and sixty-seven cents on the Friday before the announcement but opened the following Monday at two hundred twenty-two dollars and twenty-four cents, a surge of nearly thirty-five percent following the news.
This partnership emerges as OpenAI actively works to secure numerous chip alliances in its race to build extensive AI infrastructure. This initiative includes plans for five new Stargate data centers with a planned capacity of seven gigawatts. Dr. Lisa Su, the chair and CEO of AMD, stated that the company is thrilled to partner with OpenAI to deliver AI compute at a massive scale, calling the partnership a true win-win that will enable the world’s most ambitious AI buildout.
OpenAI CEO Sam Altman said the partnership is a major step in building the compute capacity required to realize AI’s full potential. This agreement is the latest in a series of recent deals for OpenAI. Last month alone, Nvidia agreed to invest up to one hundred billion dollars in OpenAI and supply the AI firm with at least ten gigawatts of compute capacity. OpenAI and Broadcom also signed a ten billion dollar deal to develop and manufacture custom AI chips. Furthermore, OpenAI confirmed the expansion of its Stargate initiative with Oracle and SoftBank. Just last week, OpenAI struck agreements with Samsung Electronics and SK Hynix to source DRAM memory chips for the Stargate project and to build data centers in South Korea.