Microsoft Builds Its Own AI Stack To Cut OpenAI Dependence

InnovationCloudMicrosoft Builds Its Own AI Stack To Cut OpenAI DependenceByJanakiram MSV,Senior Contributor.Forbes contributors publish independent expert analyses and insights. I cover emerging technologies with a focus on infrastructure and AIFollow AuthorJun 07, 2026, 10:31pm EDT--:-- / --:--This voice experience is generated by AI. Learn more.This voice experience is generated by AI. Learn more.Satya Nadella, CEO, MicrosoftMicrosoftMicrosoft spent two days at Build 2026 making the case that it no longer wants to rent the core of its AI business from anyone. The company launched seven in-house models under its MAI brand, introduced a new server processor tuned for agents, demonstrated a next-generation quantum chip and wrapped the whole thing in an agent platform that runs across Windows, Azure and GitHub. The throughline was ownership. After years of building Copilot on top of OpenAI, and more recently Anthropic, Microsoft used its developer conference in San Francisco to argue that it can supply its own intelligence, its own silicon and its own runtime.For technology leaders who have standardized on Copilot and Azure, the individual products matter less than what they signal about where the dependency now sits. Microsoft did not frame this as a break from OpenAI but as self-sufficiency, a more careful word that lets the partnership continue while it quietly builds alternatives it controls.Seven Models Trained From ScratchThe centerpiece came from Mustafa Suleyman, who runs Microsoft AI. He presented seven models spanning reasoning, coding, image generation, voice and transcription, all trained from scratch on licensed data with no distillation from rival labs. The flagship reasoning model, MAI-Thinking-1, uses a sparse mixture of experts design with roughly 35 billion active parameters and a context window of 256,000 tokens, and it is in private preview through Microsoft Foundry rather than general release. Microsoft reported that blind human raters preferred i...