Microsoft is expanding its AI capabilities with two new in-house models: MAI-Voice-1, its first natural speech generation model, and MAI-1-preview, a text-based foundation model. Developed for efficiency, MAI-Voice-1 runs on a single GPU, while MAI-1-preview utilized 15,000 Nvidia H-100 GPUs; both models aim to position Microsoft as an independent competitor in the AI space despite its reliance on OpenAI’s GPT technology for Copilot.
Want More Context? 🔎
