Microsoft AI

Overview

Microsoft AI is an organisation inside Microsoft. Its purpose, as stated in the internal memo that formed it, is to advance Copilot and Microsoft's other consumer AI products and research.^[1] Wikipedia lists its headquarters in Redmond, Washington, with offices in New York City, London, Mountain View, Zurich and elsewhere.^[2]

Origins

Microsoft AI was formed on 19 March 2024. The announcement came in an internal memo from Satya Nadella.^[1] Mustafa Suleyman, a co-founder of DeepMind and Inflection, became its executive vice president and chief executive, reporting to Nadella. Karén Simonyan, an Inflection co-founder, became its chief scientist, and several Inflection team members joined them.^[1]

At formation the Copilot, Bing and Edge teams of Mikhail Parakhin and the GenAI team of Misha Bilenko were moved under Suleyman.^[1]

On 7 April 2024 Microsoft AI opened an AI hub in London, led by Jordan Hoffmann.^[2]

Key features

In-house models. On 28 August 2025 Microsoft AI announced its first in-house models: MAI-Voice-1, a speech generation model, and MAI-1-preview, a mixture-of-experts foundation model. MAI-1-preview was trained on about 15,000 NVIDIA H100 GPUs. The same announcement said its next-generation GB200 cluster was operational.^[3]

On 13 October 2025 it announced MAI-Image-1, its first image generation model developed entirely in-house. The model debuted in the top 10 text-to-image models on LMArena.^[4]

Superintelligence work. On 6 November 2025 Suleyman announced the MAI Superintelligence Team, aimed at what the group called humanist superintelligence. This means domain-specific AI kept under human control, with medical superintelligence named as a first goal.^[5] The group cited its MAI-DxO diagnostic orchestrator as reaching 85% accuracy on New England Journal of Medicine case challenges.^[5]

Copilot reorganisation. In a 17 March 2026 reorganisation the Copilot product experience moved to a unified organisation under EVP Jacob Andreou. Suleyman's role was refocused on superintelligence and on delivering frontier models over the next five years.^[6]

Developer releases. On 2 April 2026 Microsoft AI released MAI-Transcribe-1, MAI-Voice-1 and MAI-Image-2 to developers through Microsoft Foundry and the MAI Playground.^[7]

Seven models in June 2026. On 2 June 2026 Microsoft AI launched seven new MAI models: MAI-Thinking-1, MAI-Code-1-Flash, MAI-Image-2.5, MAI-Image-2.5-Flash, MAI-Transcribe-1.5, MAI-Voice-2 and MAI-Voice-2-Flash. It described its approach as a "hill-climbing machine" that improves with more compute, better data and sharper evaluation.^[8]

Place in AI development

Microsoft AI is the part of Microsoft that was formed to carry its consumer AI products and its in-house model work. It began with Copilot and consumer products, added the MAI models from 2025, and by 2026 had a family of models for speech, images, transcription, code and reasoning. Its emphasis moved in 2026 from consumer products toward superintelligence and frontier models, with Copilot products placed under a separate leader.^[1][6][8]