AMD to Provide 6GW of Compute Capacity to OpenAI - Multi-Billion Dollar Deal

AMD and OpenAI Forge Strategic Chip Supply Partnership

AMD has entered into a multi-year agreement to supply chips to OpenAI, a collaboration projected to generate substantial revenue – potentially reaching tens of billions of dollars – for the chip manufacturer.

The agreement centers around AMD providing 6 gigawatts of computing power to OpenAI. This capacity is equivalent to the energy needed to power approximately 4.5 million homes.

Instinct GPU Series to Power OpenAI’s AI

This compute capacity will be delivered through multiple generations of AMD’s Instinct GPUs, beginning with the Instinct MI450 GPU. The initial 1 gigawatt of capacity is slated for delivery in the second half of 2026.

AMD asserts that the MI450 series will surpass the performance of comparable Nvidia offerings, specifically the Nvidia Rubin CPX, through advancements in both hardware and software.

These improvements will be developed with direct input from OpenAI. Currently, OpenAI utilizes AMD’s MI355X and MI300X series GPUs for AI inference tasks in large language models.

These existing GPUs are already recognized for their substantial memory capacity and bandwidth, making them well-suited for these demanding workloads.

Equity Stake for OpenAI

As part of this strategic alliance, AMD has granted OpenAI the option to acquire up to 160 million shares of AMD stock, representing a potential 10% equity stake.

The vesting of these shares is tied to the deployment of compute capacity, with the first tranche becoming available upon the initial 1 gigawatt deployment.

Further tranches will vest as OpenAI increases its consumption up to the full 6 gigawatts of capacity.

AMD’s stock price milestones are also linked to the equity stake, with the final tranche vesting when the stock reaches $600 per share.

As a point of reference, AMD shares closed at $164.67 on Friday, but experienced a significant increase, opening Monday at $222.24 – a rise of nearly 35% following the announcement of the deal.

Expanding AI Infrastructure

This partnership arrives as OpenAI actively seeks to establish numerous chip supply agreements to support its expanding AI infrastructure.

This includes plans for five new Stargate data centers, designed with a total capacity of 7 gigawatts.

Statements from Leadership

“We are thrilled to partner with OpenAI to deliver AI compute at massive scale,” stated Dr. Lisa Su, chair and CEO of AMD. “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.”

Sam Altman, CEO of OpenAI, commented that the partnership “is a major step in building the compute capacity to realize AI’s full potential.”

This collaboration represents the latest in a series of agreements OpenAI has secured in recent weeks to bolster its compute resources.

Recent OpenAI Partnerships

Last month, Nvidia committed to investing up to $100 billion in OpenAI and providing at least 10 gigawatts of computing power.

Additionally, OpenAI and Broadcom finalized a $10 billion deal to jointly develop and manufacture custom AI chips.

The Stargate initiative, involving Oracle and SoftBank, is also undergoing expansion.

Furthermore, OpenAI recently established agreements with Samsung Electronics and SK Hynix to procure DRAM memory chips for the Stargate project and establish data centers in South Korea.