Azure OpenAI in Foundry Models is Microsoft's service for using OpenAI models, including the GPT models, through Azure. Developers and organisations call the models through an API or the Microsoft Foundry portal to generate text, images and embeddings, to fine-tune models and to ground answers on their own data. Microsoft hosts the models in its own Azure environment; they do not interact with services operated by OpenAI itself, such as ChatGPT or the OpenAI API. According to Microsoft, prompts and completions are not available to other customers or to OpenAI, and are not used to train foundation models without permission.
Microsoft is headquartered in the United States, so the service falls under US jurisdiction, including the CLOUD Act. For standard deployments, processing stays within the geography the customer selects. Global deployments may process prompts in any geography, while a DataZone deployment in a resource located in an EU member state processes data within EU member states. Microsoft's Data Protection Addendum governs the processing.