What Project Solara Is—and Why It Matters Now
Project Solara is Microsoft’s emerging AI agent platform for enterprise devices, designed to shift AI agents from software running in browsers or apps to dedicated, managed hardware that ties together chip-level security, operating system controls, and cloud services for context-aware automation across workplace tasks and data. At Build, Microsoft described Solara as a “chip-to-cloud” platform for agent-first experiences, built on the Microsoft Device Ecosystem Platform, an Android-based foundation aimed at enterprise devices rather than consumer phones or PCs. The goal is not another smart gadget, but a common way to build and manage edge AI devices that host multiple AI agents, whether supplied by Microsoft or developed in-house. Instead of launching applications, users interact with agents that understand identity, workflows, and resources, pointing toward a post-app era where ambient AI becomes part of everyday equipment in offices, clinics, shops, and warehouses.

From Software-Only AI to Agent-First Hardware
Project Solara marks a clear move from AI agents as software features to agent-first hardware built for constant, context-rich interaction. At Build, Microsoft showed two Solara reference devices: a smart desk display and a wearable smart badge. The smart display can surface Outlook calendars, Excel data, and voice-initiated tasks, while the badge adds mobility with 5G, a touchscreen, and a side-facing camera for capturing new information on the move. These are reference designs, not commercial products, but they signal what a new class of edge AI devices could look like when AI agents are the primary interface. According to TechRepublic, Solara devices are intended as “lightweight interfaces into longer-running intelligence and action,” offloading heavy reasoning to Azure while keeping presence, identity, and input at the edge, closer to workers and workflows.

Qualcomm, Android, and the Edge AI Device Stack
Under the hood, Solara shows how Microsoft is rethinking the device stack for AI agents. The platform runs on Microsoft Device Ecosystem Platform, a fork of Android Open Source Project tailored for enterprise devices, not consumer Android phones. That gives hardware makers a familiar base while tying into Microsoft tools, identity, and management. Qualcomm and MediaTek are the first silicon partners for Solara concept designs, bringing hardware acceleration to agent workloads so that voice, vision, and on-device reasoning feel responsive without relying wholly on the cloud. Satya Nadella highlighted that Microsoft is not only improving AI models but “rethinking the entire computing stack, from devices and operating systems to agents and enterprise workflows.” In practice, that means chip-to-cloud design: approved chipsets, standardized security features, and cloud-backed agents aligned with Azure, Copilot, and other Microsoft services.
Identity, Privacy, and IT Management for Agent-Enabled Devices
For enterprise IT, Solara is less about flashy gadgets and more about a new category of managed endpoints. Solara-style devices may need access to microphones, cameras, biometrics, transcripts, and sensitive business data. The reference designs therefore include Entra ID sign-in, Windows Hello for Business, privacy switches, and support for Intune management, signaling that Microsoft expects these to be treated like any other managed endpoint. Device identity, policy enforcement, and auditability become central: admins must decide which agents can run, what data they can see, and how recordings or logs are retained. Healthcare pilots around Dragon Copilot illustrate the stakes, where clinicians rely on agents for documentation and follow-up tasks while regulators expect clear consent, record-keeping, and boundaries on data use. Solara brings these questions out of browsers and PCs and into badges, wall displays, and other edge AI devices.
Preparing for Ambient AI Across Enterprise Devices
Solara points toward a future of ambient AI, where AI agents live on specialized edge AI devices rather than only in PCs or cloud-hosted chat windows. Microsoft describes Solara as designed for a “multiple-agent world,” with plans for an agent dispatcher and task manager so users are not locked into a single, dominant agent. Pilots with Target, CVS Health, Best Buy, Levi’s, and others will test how these agent-first hardware concepts work in frontline contexts like retail, logistics, and field operations. For now, Microsoft has not announced general availability or pricing, so Solara should be tracked, not bought. IT leaders can use this window to review policies for always-on sensors, clarify data retention and consent rules for AI transcripts, and strengthen identity governance—so when AI agent platforms like Solara arrive at scale, the enterprise is ready to manage them as first-class, policy-bound devices.






