The agreement between the companies, set until 2030, gives Microsoft the right of first choice when OpenAI needs computing power for training and operating its models. This means that Microsoft can offer its services first before OpenAI turns to competitors.
This step follows an exception made last year when OpenAI was allowed to use Oracle’s cloud capabilities once. Now, the companies have secured the possibility of broad cooperation between OpenAI and other providers on a permanent basis.
Additionally, OpenAI has been granted permission to develop its own cloud capabilities, which will be used for research and model training.
