Microsoft to host AI models from Meta, xAI, and others in own data centres

Microsoft has increased Azure AI models to over 1,900 with new partnerships. (Photo: Shutterstock)

Microsoft has announced plans to host AI models developed by Meta Platforms, xAI, and European firms Black Forest Labs and Mistral in its data centres. This development was revealed at the Build software developer conference in Seattle, marking a strategic shift in Microsoft’s AI approach, as per by report in Reuters. Redmond aims to provide these models with the same reliability assurances as those hosted for OpenAI.

“That’s just a game-changer in terms of how you think about models and model provisioning,” said CEO Satya Nadella. “It’s exciting for us as developers to be able to mix and match and use them all.”

Microsoft will also offer xAI’s Grok 3 and Grok 3 mini models, alongside Meta’s Llama models and solutions from Mistral and Black Forest Labs, increasing the total AI models available to Azure clients to over 1,900. These models will operate within Microsoft’s data centres to maintain consistent availability. Redmond plans to introduce more popular models soon.

Microsoft intends to remain neutral in the AI market by broadening its partnerships and focusing on economical sales strategies.

Meanwhile, GitHub and Microsoft have unveiled a new coding agent for GitHub Copilot. The new agent operates when a GitHub issue is assigned or a prompt is received in Visual Studio Code. This creates a secure development environment via GitHub Actions, with the agent logging details and submitting draft pull requests for review. Existing security measures, such as branch protections, remain effective, requiring human approval for pull requests before any CI/CD workflows are executed.

At the conference, Microsoft also presented Azure Foundry, a service allowing businesses to create customised AI agents for internal tasks. Asha Sharma, Microsoft’s VP for AI platforms, noted that these agents would likely utilise multiple AI models.

Furthermore, Microsoft is developing digital identifiers for AI agents similar to those for human employees. Other new offerings from Microsoft include the Model Leaderboard, which ranks the best-performing AI models across various categories and tasks. Additionally, a new tool called Model Router can help in picking the most suitable model for a particular query or task in real-time.

FTC investigates Microsoft over antitrust issues

It can be noted that in March 2025, Bloomberg News reported that the US Federal Trade Commission (FTC) is advancing an antitrust investigation into Microsoft, initiated at the end of the Biden administration. The FTC is examining allegations of Microsoft’s potential market power misuse in productivity software, particularly through licensing terms that could deter customers from switching from Azure to rival platforms. The FTC has requested information on Microsoft’s data centres, its efforts to meet customer demand, and expected licensing changes. The report also notes scrutiny on Microsoft’s funding reduction for its AI projects post-OpenAI agreement and its cybersecurity and AI product practices.

Read more: OpenAI to rework Microsoft partnership, reduce revenue share