According to Bloomberg, Microsoft is sharpening its stance against the very AI labs it partners with. The company has reportedly given salespeople talking points to steer enterprise customers away from Anthropic and OpenAI, emphasizing cost and security, and is said to be preparing a multi-model security product that routes tasks across Microsoft, OpenAI and Anthropic models depending on which performs best. Separately, Bloomberg reported Microsoft has begun swapping in its own MAI models for some prompts in apps like Excel and Outlook. The moves show Microsoft trying to cut its dependence on outside models while still reselling them, a delicate balance given its deep ties to OpenAI. Watch how the labs respond as their largest distributor becomes a rival.