Microsoft showcased its biggest push yet into artificial intelligence (AI) at its annual Build developer conference, unveiling new AI models, developer tools, Windows capabilities and quantum computing advances designed to accelerate the next generation of intelligent applications.

Leading the announcements was the Microsoft Agent Platform, which allows developers to build AI agents in GitHub, deploy them through Microsoft Foundry and optimize them using multiple AI models while maintaining enterprise-grade security and governance. The company also introduced Microsoft IQ, a new context layer that grounds AI agents in enterprise data, workplace knowledge and live web information.

Microsoft expanded its in-house AI portfolio with MAI-Thinking-1, a 35-billion-parameter reasoning model built for complex tasks and code generation, alongside new image, voice, transcription and coding models under the broader MAI family. The company also debuted a preview of a redesigned GitHub Copilot app that transforms the coding assistant into a full agentic development environment.

On the hardware side, Microsoft introduced the Surface RTX Spark Dev Box, an AI-focused developer machine powered by NVIDIA RTX Spark capable of running large language models locally. Windows also gained new AI capabilities, including Microsoft Execution Containers and OpenClaw support for securely running autonomous AI agents.

The company further reinforced its AI safety efforts by open-sourcing ASSERT, a policy-driven evaluation framework, and the Agent Control Specification, which standardizes security controls for AI agents across different platforms.

Microsoft also highlighted progress in quantum computing with Majorana 2, its latest quantum chip that delivers a 1,000-fold improvement in reliability over its predecessor and advances the company’s goal of building a scalable quantum computer by 2029. Complementing the hardware announcement, Microsoft said its AI-powered research platform, Microsoft Discovery, is now generally available to help scientists accelerate research through agent-assisted workflows.