AMD to supply 6GW of compute capacity to OpenAI in chip deal worth tens of billions

AMD has signed a multi-year chip supply deal with OpenAI that could generate tens of billions in revenue for the chipmaker, helping accelerate its momentum in the AI industry. 

AMD has agreed to supply 6 gigawatts of compute capacity to OpenAI — enough to power up to 4.5 million homes — across multiple generations of its Instinct GPUs, starting with the Instinct MI450 GPU. OpenAI will receive the first gigawatt of capacity in the second half of 2026, when the new chip is scheduled for deployment. 

AMD claims the MI450 series will outperform Nvidia’s comparable offerings (Nvidia Rubin CPX) through hardware and software improvements, many of which will be made with OpenAI’s input. Its current MI355X and MI300X series GPUs, currently used in some workloads for OpenAI, are already strong for AI inference in large language models due to their large memory capacity and bandwidth.  

As part of the agreement, AMD has given OpenAI the option to buy up to 160 million shares of AMD stock, which amounts to a 10% stake. The first tranche will vest with the initial 1 gigawatt deployment, and additional tranches will vest as OpenAI buys up to the total 6 gigawatts, AMD said. 

OpenAI’s stake will also be directly tied to increasing AMD’s stock price milestones, with the final tranche vesting when the stock reaches $600 per share. For context, AMD shares closed at $164.67 on Friday, but started off Monday at $222.24, up nearly 35% following news of the deal.

The deal comes as OpenAI works to secure as many chip partnerships as possible in its race to build out AI infrastructure, including five new Stargate data centers with a planned capacity of 7 gigawatts.  

“We are thrilled to partner with OpenAI to deliver AI compute at massive scale,” said Dr. Lisa Su, chair and CEO of AMD, in a statement. “This partnership brings the best of AMD and OpenAI together to create a true win-win enabling the world’s most ambitious AI buildout and advancing the entire AI ecosystem.” 

OpenAI CEO Sam Altman said in a statement that the partnership is “a major step in building the compute capacity to realize AI’s full potential.” This is the latest of several deals the AI giant has made in recent weeks to secure compute capacity.  

Last month alone, Nvidia agreed to invest up to $100 billion in OpenAI and supply the AI firm with at least 10 gigawatts; OpenAI and Broadcom also signed a $10 billion deal to develop and manufacture custom AI chips; and OpenAI said its Stargate initiative with Oracle and SoftBank would be expanding.  

Last week, OpenAI struck agreements with Samsung Electronics and SK Hynix to source DRAM memory chips for the Stargate project and build data centers in South Korea.