• Microsoft creates in-house AI models, called MAI.
• Move reflects Microsoft’s strategy to reduce dependence on OpenAI.

Microsoft is working on in-house AI models that could compete with industry leaders, including its long-term partner OpenAI, according to a source familiar with the matter.

The company has been testing a family of models known as MAI, which apparently perform on par with leading AI models from OpenAI and Anthropic. Microsoft is examining how the models could improve its existing products, like the Copilot-branded AI assistants. Copilot is intended to manage user queries and provide suggestions for tasks like document editing, software development, and conference calls.

Expanding AI capabilities across Microsoft products

Beyond Copilot, Microsoft is considering ways to integrate MAI into other products and services:

• Microsoft Teams – MAI could enhance real-time transcription, language translation, and automated meeting summaries.
• Azure Cloud Services – MAI could be used to automate enterprise-level tasks, like customer service, data analysis, and infrastructure management.
• LinkedIn – MAI might improve job recommendations and recruitment.

By embedding MAI into these platforms, Microsoft could offer what it terms a more connected and intelligent ecosystem, and strengthen its position in the AI market.

Developing reasoning models

Microsoft is also working on reasoning models that can handle more difficult problem-solving tasks and imitate human-like decision-making. OpenAI, Anthropic, and Alphabet are also developing similar models. Microsoft incorporated OpenAI’s o1 reasoning model into Copilot last month as part of its ongoing collaboration with OpenAI (of which it owns 49%). A Microsoft spokesperson said the company continues to use a mix of models from OpenAI, Microsoft AI, and open-source sources to support its products.

Reducing dependence on OpenAI

The development of MAI reflects Microsoft’s broader effort to reduce its reliance on OpenAI. Microsoft has invested about $13 billion in its partner since 2019. The partnership granted OpenAI access to Microsoft’s Azure cloud platform for AI research and development at a reduced rate. The two companies renegotiated their agreement in January, allowing OpenAI to use cloud services from competitors unless Microsoft claims the business for itself. The updated deal runs until 2030.

Microsoft Chief Financial Officer Amy Hood recently discussed the partnership at a Morgan Stanley conference. “We’re both successful when each of us are successful,” Hood said. “So as you go through that process, I do think everybody’s planning for what happens for a decade, or two decades. And that’s important for both of us to do.”

Since the partnership began, OpenAI’s models have been integrated into several Microsoft products, including Office, GitHub Copilot, and Bing Search. Microsoft’s AI infrastructure is hosted on Azure, and the company has collaborated with OpenAI on AI supercomputing and LLMs.

“We feel great about having leading models from OpenAI, we’re still incredibly proud of that,” Hood previously said. “But we also have other models, including ones we build, to make sure that there’s choice.”

Expanding AI model options

Alongside OpenAI’s models and MAI, Microsoft has developed a set of smaller in-house models called Phi. The company has also tested AI models from other providers, including Anthropic, DeepSeek, Meta, and xAI, to assess their performance in the Copilot framework. Creating MAI reflects Microsoft’s strategy of diversifying its AI capabilities. Expanding its model portfolio helps the firm to provide more flexibility and reduce dependency on any single AI provider.

With this approach, Microsoft could offer tailored artificial intelligence solutions for different business segments. Microsoft’s strategy to combine in-house models with third-party solutions reflects its broader plan to balance independence with strategic partnerships. OpenAI remains a key partner, but developing MAI models positions Microsoft to respond to shifts in the AI market and offer more adaptable AI solutions.