Microsoft has been a significant player in the AI Revolution since day one. However, until now, the company has primarily depended on the capabilities of OpenAI’s large language models (LLMs) to power its next-generation AI tools and platforms. This is about to change, as the company announces a first pair of in-house models developed by the Microsoft AI (MAI) team: MAI-Voice-1 and MAI-1-preview.
What to Expect
A recent Microsoft blog post explains that the MAI-1 models have been purpose-built to enable the MAI team to fully meet its ambitious goals, namely “creating applied AI as a platform for category defining and deeply trusted products that understand each of our unique needs.”
The first model to be released is MAI-Voice-1. It is a natural speech generation model available through Copilot Daily and Podcasts, as well as a user experience ready to test out in Copilot Labs. Voice is compatible in both single-speaker and multi-speaker settings, delivering highly expressive audio outputs.
I tried the feature out, and the nuances between emotions, the continuation of text from simple prompts, and the adaptability are outstanding. Remarkably, the model can generate a minute of audio output in under a second with just a single GPU.
Microsoft has also begun public testing of MAI-1-preview on LMArena, a public, web-based LLM evaluation platform. The model is Microsoft’s “first foundation model trained end-to-end” and a mixture-of-experts model that the company pre-trained and post-trained on approximately 15,000 NVIDIA H100 GPUs.
Microsoft has announced that it will be “rolling MAI-1-preview out for certain text use cases within Copilot over the coming weeks to learn and improve from user feedback.”
AI Agent & Copilot Summit is an AI-first event to define opportunities, impact, and outcomes with Microsoft Copilot and agents. Building on its 2025 success, the 2026 event takes place March 17-19 in San Diego. Get more details.
A Souring Relationship
The shift away from OpenAI, or at least the burgeoning gap between the two companies, has been ongoing since the start of the year. Here’s a timeline of events as covered by Cloud Wars:
January 2025: “Microsoft Diversifies LLM Options for 365 Copilot, Expanding Focus Beyond” OpenAI — Microsoft is ensuring it can respond to consumer demand by giving its customers what they need: flexibility and choices, while continuing to benefit from one of the most important relationships in the technology industry.
May 2025: “Microsoft and OpenAI Could Revise Partnership Terms Ahead of Potential OpenAI IPO” — Ultimately, OpenAI is becoming a potential rival to Microsoft. However, predicting how this will manifest is difficult, and while OpenAI has dominated the GenAI sector and the LLM market, Microsoft has consistently played the big brother role.
June 2025: “OpenAI Partners with Google Cloud Amid Soaring AI Demand, Reshaping Ties with Microsoft” — Along with the Google Cloud deal, OpenAI has also become a key partner in the $500 billion Stargate Project, marking a significant shift in its strategic alliances and potentially its future direction.
August 2025: “Larry Ellison and Oracle Beat Microsoft for Largest Tech Contract Ever: $100-Billion OpenAI Stargate Deal” — Oracle’s name and not Microsoft’s is on the largest tech contract ever created — $100 billion over three years — and in fiscal 2028 OpenAI will begin paying Oracle the equivalent of $82.2 million every single day to drive unprecedented innovation in the AI Revolution.
However, it’s important not to get bogged down by the Microsoft/OpenAI breakup, because the fundamentals and opportunities presented by Microsoft’s entry into the foundation model market are big news.
I’ll conclude with a quote from the Microsoft blog I referenced above to illustrate this point: “We have big ambitions for where we go next. Not only will we pursue further advances here, but we believe that orchestrating a range of specialized models serving different user intents and use cases will unlock immense value.”
Ask Cloud Wars AI Agent about this analysis
