Microsoft has unveiled Project Solara, an AOSP-based platform that replaces traditional apps with AI agents, signalling a shift towards task-driven computing and dynamically generated user interfaces.
Microsoft has unveiled Project Solara at Build 2026, positioning the Android Open Source Project (AOSP) as the foundation for a new generation of agent-first devices designed to “run AI agents instead of traditional applications.”
The platform aims to replace the app-centric computing model with a task-driven approach, where AI agents interpret user intent, execute workflows across systems, and deliver results without requiring users to open individual applications.
Built as a chip-to-cloud architecture, Solara combines AOSP with Microsoft’s enterprise stack, including Entra ID, Intune, Copilot-powered agents, and security and management tools. The system spans local hardware and cloud infrastructure, enabling agents to access enterprise systems, cloud data, and device inputs while coordinating tasks in the background.
A key feature is Microsoft’s “just-in-time UI” model, which dynamically generates interfaces based on user intent, context, and device type. Rather than relying on fixed layouts, the same task can produce different outputs on a wearable device and a larger display, potentially reducing the need for developers to create multiple screen-specific interfaces.
Microsoft also showcased two reference designs: a wearable badge featuring a touchscreen, camera, microphone array, and biometric authentication, and a desk companion device that can function as a cloud PC client when connected to an external display. Notably, Solara-powered devices do not feature traditional applications, app stores, or desktop interfaces.
The platform remains in an early stage, with pilot deployments reportedly underway at Best Buy, CVS Health, Levi’s, and Target. Its success will depend on the maturity of AI agents and whether enterprises are willing to embrace a fundamental shift away from established software ecosystems.














































































