Microsoft has reportedly begun replacing AI models from OpenAI and Anthropic with its own in-house MAI (Microsoft AI) models across key Microsoft 365 applications, including Excel and Outlook. The move marks a significant shift in Microsoft’s AI strategy as it looks to reduce operational costs and lessen its dependence on third-party AI providers.
According to reports, Microsoft’s internally developed models are already handling tens of thousands of AI prompts every week within these productivity applications, signalling the company’s growing confidence in its own AI capabilities.
Cost Reduction Drives The Transition
Microsoft’s investment in artificial intelligence has grown rapidly over the past few years, resulting in billions of dollars in infrastructure and cloud computing expenses. Running premium AI models from OpenAI and Anthropic at scale has become increasingly expensive, prompting Microsoft to optimise costs by deploying its own MAI models wherever possible.
The company is reportedly routing AI requests in Excel and Outlook through its proprietary models instead of external ones, allowing it to reduce AI inference costs while maintaining similar user experiences.
MAI Models Expand Across Microsoft Products
The rollout currently focuses on Excel and Outlook, but Microsoft is expected to expand the use of its MAI models across more products in the Microsoft ecosystem. Reports suggest that future deployments could include GitHub Copilot, Microsoft Teams, and other AI-powered services.
Microsoft introduced several new MAI models earlier this year as part of its broader ambition to build a comprehensive AI platform capable of competing with leading foundation model providers.
Partnership With OpenAI Remains Intact
Despite the transition, Microsoft is not ending its relationship with OpenAI. Instead, the company appears to be adopting a hybrid strategy where it can choose between in-house and third-party models depending on cost, performance, and specific application requirements.
This flexibility allows Microsoft to reduce reliance on any single AI provider while continuing to offer customers access to advanced AI capabilities where needed. Industry analysts view the move as part of a broader trend towards model diversification rather than exclusive partnerships.
A New Phase In Microsoft’s AI Strategy
Microsoft’s decision reflects the evolving economics of generative AI. As AI adoption grows across enterprise software, technology companies are increasingly seeking greater control over both performance and operating costs.
By deploying its own MAI models across high-volume applications like Excel and Outlook, Microsoft aims to improve profitability while strengthening its long-term AI independence. The strategy could also pave the way for more customised AI experiences across its productivity suite and enterprise offerings.
Summary (60 Words)
Microsoft has started replacing OpenAI and Anthropic models with its in-house MAI models in Excel and Outlook to reduce AI operating costs and decrease reliance on external providers. The proprietary models are already processing thousands of AI requests weekly, with plans to expand across additional Microsoft products while maintaining a flexible, multi-model AI strategy.
