Microsoft’s Declaration of Independence: The New MAI Models Challenging OpenAI and Google

Microsoft announced the MAI model family—MAI-Transcribe-1 (speech-to-text), MAI-Voice-1 (text-to-speech), and MAI-Image-2 (image generation)—claiming superior benchmarks, lower GPU usage, aggressive pricing, and enterprise-focused compliance; leadership framed the effort as a move toward AI self-sufficiency after renegotiating licensing constraints with OpenAI and emphasized that small, highly skilled teams produced these models.