OpenAI has once again expanded its AI computing territory. This time, its partner is no longer NVIDIA, but another chip manufacturer AMD.A sizable agreementOpenAI will use AMD's Instinct MI450 GPU to build up to 6 GW (gigawatt) of AI computing infrastructure, and AMD expects to earn "tens of billions of dollars" from it.
Under this agreement, AMD will become OpenAI's "core strategic computing partner," assisting in the deployment of large-scale AI systems. OpenAI plans to use AMD's Instinct series GPUs, with the initial rollout of Instinct MI450 GPUs expected to begin in the second half of 2026, marking the official launch of the first phase of a 1 GW AI computing deployment.
OpenAI CEO Sam Altman said that AMD's development position in the field of high-performance chips will help OpenAI advance technological innovation more quickly, "allowing the benefits of AI to benefit users around the world more quickly."
In addition to technical collaboration, the agreement also includes a potential equity exchange. OpenAI will have the opportunity to purchase up to 1.6 million AMD shares at a symbolic price of $0.01 per share, with the option to acquire shares incrementally as milestones are achieved. Ultimately, OpenAI could acquire approximately 10% of AMD's shares, demonstrating that the relationship between the two companies will extend beyond partnership to encompass a long-term strategic alliance.
The timing of this deal also coincides with OpenAI's recent frequent alliances with major chip suppliers. A few weeks ago, OpenAI announced a total partnership with NVIDIA.Long-term investment plan of up to US$1000 billion, used to build an AI data center of at least 10 GW in scale, which is also scheduled to be launched gradually from the second half of 2026.
At the same time, OpenAI continues to deepen its partnership with Microsoft, which has invested more than $130 billion and holds a 49% profit share in it. On the other hand, NVIDIA has recently made aInvest $50 billion, thereby promoting cooperation between the two parties in AI acceleration computing and x86 architecture integration, and even Intel building a customized AI platform and CPU for it.
Overall, this wave of strategic alliances between chip and AI giants is reshaping the global AI computing ecosystem. OpenAI's collaboration with AMD not only reveals a new landscape of GPU competition but also signals that AI infrastructure is transitioning from a single supply chain to a more diversified and hyperscaled stage.
