OpenAI has entered into a multi-year partnership with AMD that will see the chipmaker supply up to 6 gigawatts of AI computing power using its upcoming Instinct MI450 GPUs. The first facility under this deal is expected to go live in the second half of 2026 with a capacity of one gigawatt. The agreement highlights the growing scale of OpenAI’s infrastructure expansion and the company’s efforts to secure as much compute power as possible to meet its rising AI demands.
Sam Altman, CEO of OpenAI, said the partnership will help the company build enough AI infrastructure to support its growing computational needs. The deal positions AMD as a stronger competitor to Nvidia in the AI hardware market, although Nvidia still leads in overall scale and revenue.
However, this is not the only major infrastructure announcement from OpenAI. Earlier this year, OpenAI signed a 10-gigawatt GPU deployment deal with Nvidia, with Nvidia committing up to $100 billion in investment. Apart from that, the company has also teamed up with Oracle and SoftBank under a project known as Stargate to build up to 10 gigawatts of AI data center capacity in the U.S. by the end of 2025.
Executives at OpenAI have repeatedly said that the company faces “compute scarcity” due to the rising demand for its AI models. President Greg Brockman recently stated that OpenAI needs “as much computing power as we can possibly get,” pointing to the company’s aggressive infrastructure expansion across multiple chip vendors and data center operators.
Beyond AMD, Nvidia, and Oracle, OpenAI has also deepened its partnerships with CoreWeave and other cloud providers to secure additional GPU capacity. These projects require massive energy resources, often compared to the electricity usage of entire cities, and OpenAI is working on integrating renewable energy sources like solar power, refurbished gas turbines, and small modular nuclear reactors to ensure long-term sustainability.
