Microsoft ‘has begun testing out models from xAI, Meta and DeepSeek as potential OpenAI replacements in Copilot’

Just a day after reports surfaced that Microsoft is planning a future without OpenAI, Reuters revealed on Friday that the Redmond-based tech giant, a major investor in OpenAI, has started testing models from xAI, Meta, and DeepSeek as possible alternatives for Copilot.

“The Redmond, Washington-based company, a major backer of OpenAI, has begun testing out models from xAI, Meta and DeepSeek as potential OpenAI replacements in Copilot, Reuters reported, citing The Information.

Microsoft is Developing in-house AI Reasoning Models

In addition to testing new AI models, the report added that Microsoft is also developing its own artificial intelligence reasoning models. These models could be offered to developers as an alternative to OpenAI’s technology, the report stated.

“Microsoft is developing in-house artificial intelligence reasoning models to compete with OpenAI and may sell them to developers, The Information reported on Friday, citing a person involved in this initiative.”

The move signals Microsoft’s efforts to reduce its reliance on OpenAI, despite an early partnership that positioned the company ahead of its rivals in the AI space.

In December, Reuters reported that Microsoft was working to integrate both in-house and third-party AI models into Microsoft 365 Copilot. This shift is aimed at expanding its AI capabilities beyond OpenAI’s technology while also managing costs.

When 365 Copilot was introduced in 2023, a key selling point was its integration with OpenAI’s GPT-4 model. However, The Information now reports that Microsoft’s AI division, led by Mustafa Suleyman, has trained a set of models called MAI. These models reportedly perform on par with OpenAI and Anthropic’s leading models on widely used benchmarks.

Microsoft is also developing reasoning models using chain-of-thought techniques, which allow AI to break down complex problems into intermediate steps before generating an answer. This approach could put Microsoft’s models in direct competition with OpenAI.

Suleyman’s team is already experimenting with using the MAI models in place of OpenAI’s technology in Copilot. These models are significantly larger than Microsoft’s earlier Phi models, according to the report.

Microsoft is exploring the possibility of making the MAI models available later this year as an API, which would allow developers to integrate them into their own applications.

This isn’t a new idea inside Microsoft. Back in December 2024, Reuters reported that the company had already begun integrating its in-house models, like Phi-4, into Microsoft 365 Copilot to speed things up and bring down costs. The partnership with OpenAI has been lucrative—Microsoft gets a share of the revenue—but it hasn’t been without strain. OpenAI’s projected $5 billion loss in 2024, flagged by The New York Times, only reinforced Microsoft’s desire to have its own AI playbook. Suleyman’s hiring was a clear sign of that shift.

Will Microsoft Replace OpenAI’s ChatGPT with xAI’s Grok?

While Suleyman’s team has been testing alternatives from companies like xAI, DeepSeek, and Meta, the online chatter today reflects a mix of skepticism and intrigue. Some are questioning whether Microsoft can truly break free when OpenAI’s tech is so embedded in its ecosystem.

If Microsoft pulls this off, the benefits are obvious: cheaper, faster AI services and a chance to stake out its own ground in a competitive market. But between contractual obligations, technical dependencies, and OpenAI’s ongoing innovation, Suleyman’s mission is more of a long grind than a quick win. His team is making moves—one model swap at a time. Can Microsoft fully wean itself from OpenAI? That’s the billion-dollar question.

What This Means for OpenAI and the AI Industry

Microsoft testing alternatives to OpenAI marks a significant shift in AI. Given its deep investment in OpenAI, this move could shake up the industry.

For OpenAI, it raises questions about stability and reliance on a single major backer. For Microsoft, it’s a way to take greater control of its AI strategy and cut costs.

Some will see this as a logical step—companies spread their risks rather than relying too much on one provider. Others may see it as a sign that the Microsoft-OpenAI partnership isn’t as solid as it once seemed.