The real story, per TechCrunch AI's reporting, is not this new 'preferred' label but the underlying tension it seeks to paper over. Declaring a model 'preferred' is a marketing term, not a technical or contractual one, and does little to disprove the core claim that Microsoft is actively pursuing cost reduction by using its own MAI models. In our view, this announcement looks like a classic 'the lady doth protest too much' moment, designed more for enterprise customers and investors than for the product roadmap. The strategic dependency here is asymmetric: Microsoft can afford to experiment with its own stack; OpenAI cannot afford to lose its flagship enterprise deployment.