Microsoft: Less OpenAI, More Own AI in Office and Outlook

Microsoft is starting to replace OpenAI and Anthropic's artificial intelligence models with its own. The reason is simple and very down-to-earth: to cut costs. Spending less to run AI in its own apps is the primary goal of this strategic move.

Until recently, the Redmond company heavily relied on external partners for advanced AI capabilities. Now, with its own internal models ready, it aims for greater autonomy. It's a bit like switching from external catering to in-house cooking, hoping to save money without sacrificing dish quality. Or at least, that's the theory. Microsoft Corp. began replacing OpenAI and Anthropic models with its own in some software products as of July 7, 2026.

The first apps to see the AI "brain" change will be the ones we use most, like Excel and Outlook. Imagine writing assistance or data analysis features that, instead of going through OpenAI's servers, will use technology developed directly by Microsoft.

This transition could lead to tighter integration and, perhaps, optimized performance. Or maybe not, and no one will notice the difference. The point is, Microsoft wants more control and fewer expenses. Who wouldn't, after all? It’s like when your trusted supplier raises prices and you decide to produce in-house. The hope is that the final result is just as good, if not better.

Not really, but it's a sign the wind is changing. Microsoft has invested billions in OpenAI, but every company's goal is always to maximize profits. If they can do the same thing, or almost, with their own resources and spend less, why shouldn't they?

The relationship continues, but with more pragmatism. Microsoft will keep using OpenAI's models for more complex applications or where its own models aren't quite up to par. But for the daily "grunt work" in Office, well, it'll do it itself. It’s a textbook move to reduce dependency and, crucially, those hefty external AI subscription bills.