Advanced Micro Devices (AMD) has entered a significant partnership with OpenAI involving a 6-gigawatt commitment to fuel OpenAI’s AI infrastructure using multiple generations of AMD Instinct GPUs. Initial deployments of the MI450 GPUs are expected in the latter half of next year.

OpenAI has named AMD as a strategic partner for compute resources, signaling plans to introduce a series of AI-centric rack solutions and ensure support for future AMD GPU iterations. This agreement follows OpenAI’s recent collaboration with Nvidia in a substantial chip deal aimed at procuring 10 gigawatts worth of Nvidia systems, linked to a potential $100 billion investment in datacenters.

As part of the AMD agreement, OpenAI has been issued a warrant for up to 160 million shares of AMD stock, conditional on meeting certain performance and share-price milestones.

Dr. Lisa Su, AMD’s CEO, expressed her enthusiasm for the venture, emphasizing the mutual benefits brought forth by the collaboration which aims to expand the global AI ecosystem.

Following the announcement, AMD experienced a notable surge in its stock value. However, some industry observers remain skeptical, likening it to excessive optimism in the face of unproven profitability from emerging technologies.

OpenAI CEO Sam Altman highlighted the partnership as critical in scaling compute power to fully harness AI’s capabilities, amidst OpenAI’s rapid expansion and strategic deals with technology giants like Samsung and SK Hynix. While valued highly in recent market estimations, OpenAI is not projected to achieve positive cash flow until 2029.