In a move signaling Microsoft’s evolving strategy in the AI landscape, the tech giant is developing its own artificial intelligence (AI) reasoning models, codenamed ‘MAI.’ This internal initiative aims to reduce its dependency on OpenAI, the current leader in the field, and enhance its AI capabilities.
According to a report by The Information, these ‘MAI’ models are designed to tackle complex reasoning and problem-solving tasks. Microsoft’s ambition is for the MAI models to match or even exceed the performance of leading models from OpenAI. The Redmond-based company is actively experimenting with integrating the MAI models into its products, including Microsoft 365 Copilot. Furthermore, Microsoft is considering releasing these models as an application programming interface (API) later this year. This would empower external developers to incorporate Microsoft’s AI capabilities into their applications, offering potential advantages in cost and flexibility compared to exclusive reliance on OpenAI.
This development doesn’t come as a surprise. Reports from December 2024 suggested that despite a substantial investment of $13.75 billion in OpenAI since 2019, Microsoft has been exploring alternatives to the ChatGPT maker’s AI models, particularly for its 365 Copilot. Initially, Microsoft 365 Copilot, launched in 2023, prominently featured OpenAI’s GPT-4 model. However, the scenario appears to be shifting, potentially for several key reasons: to avoid dependency on a single provider, to reduce costs, and to improve speed.
The ‘MAI’ initiative is being led by Microsoft’s AI division under Mustafa Suleyman. The team has already completed the training of a family of models that perform nearly as well as OpenAI’s leading models. These reasoning models utilize ‘chain-of-thought’ techniques to solve intricate problems by generating answers with intermediate reasoning steps.
The latest reports also indicate Microsoft is testing AI models from xAI, Meta, and DeepSeek as potential replacements for OpenAI’s technology within Copilot. This strategic shift coincides with Microsoft’s recent introduction of Copilot AI features to its Microsoft 365 suite for Personal and Family subscription plans, expanding access that was formerly exclusive to the separate Copilot Pro subscription at an additional monthly cost.
Earlier this year, Microsoft modified its partnership with OpenAI, moving away from its previous exclusive cloud provider role for OpenAI’s computing needs. Under the revised agreement, Microsoft retains a ‘right of first refusal’ (ROFR), providing the first opportunity to supply cloud resources to OpenAI. However, OpenAI now has the option to seek services from other cloud providers, such as Oracle.
The timing of this evolution in the Microsoft-OpenAI relationship overlaps with the unveiling of the Stargate Project. This collaboration between OpenAI, Oracle, SoftBank, and Nvidia is expected to inject up to $500 billion into developing the necessary infrastructure to support next-generation AI models in the United States. Concurrently, Microsoft plans to invest $80 billion in AI-focused data centers and cloud infrastructure in fiscal year 2025, a portion of which will be allocated to India.
