Microsoft Launches Proprietary AI Models to Cut Dependence on OpenAI Services

Microsoft announced at its annual Build conference that it has developed two new large‑language AI models designed to run on Azure, marking the company’s first major step toward offering alternatives to OpenAI’s services.

The new models are positioned to handle enterprise workloads such as code generation, document summarization and customer‑service chat, and Microsoft said they will be available to developers through Azure AI Studio starting later this quarter.

Azure platform will host the models on dedicated hardware clusters, and Microsoft highlighted tighter integration with its existing security and compliance tools, aiming to give businesses greater control over data residency and model behavior.

Industry analysts view the move as a strategic push to diversify the AI supply chain, noting that reducing reliance on a single external provider could lower costs for large‑scale adopters and spur competition in the cloud AI market.