𝗠𝗶𝗰𝗿𝗼𝘀𝗼𝗳𝘁 𝗜𝘀 𝗕𝘂𝗶𝗹𝗱𝗶𝗻𝗴 𝗜𝘁𝘀 𝗢𝘄𝗻 𝗔𝗜 𝗙𝘂𝘁𝘂𝗿𝗲
Microsoft just changed its AI strategy.
For years, Microsoft relied on OpenAI. They invested billions and integrated GPT into almost every product. But at Build 2026, Microsoft announced seven new in-house AI models.
Why the shift?
Microsoft realized that owning the model is better than renting it.
The partnership with OpenAI changed in late 2025. OpenAI became a public benefit corporation and opened doors to other cloud providers. Microsoft could no longer rely on OpenAI as its exclusive partner.
Now, Microsoft is moving from consuming models to building them.
The new MAI model family includes:
• MAI-Thinking-1: A flagship reasoning model. • MAI-Code-1-Flash: A fast model for coding. • MAI-Image-2.5: For image generation and editing. • MAI-Transcribe-1.5: For transcription. • MAI-Voice-2: For speech tasks.
These models offer three main advantages for you:
Lower Costs Microsoft now runs its own models on its own chips, like the Maia 200. This vertical integration reduces expenses. They can offer high performance at a fraction of the previous cost.
Better Data Privacy Microsoft trained these models from scratch using commercially licensed data. They did not use distillation from other companies. This gives enterprises more confidence in data lineage and IP safety.
More Choice You are no longer locked into one provider. You can use OpenAI for complex reasoning, but use MAI models for tasks where speed and cost matter most.
Microsoft is not leaving OpenAI. The partnership continues through 2032. However, Microsoft is no longer dependent on them.
The era of single-vendor AI is ending. The future belongs to companies that own their entire stack.