Peter Steinberger demonstrated OpenClaw as a native Windows application during the Microsoft Build conference in Seattle on June 2, 2026 [1].

The demonstration highlights a shift toward making powerful AI agents viable for corporate use by addressing the security vulnerabilities often associated with autonomous software. By utilizing Microsoft Execution Containers, the integration aims to isolate the AI's actions from the core operating system.

Steinberger, the creator of OpenClaw, presented the tool as a native app to show how it can function within enterprise environments. The use of Execution Containers is intended to provide a secure layer that allows the software to operate without compromising the host machine's integrity [1].

"You can totally run OpenClaw inside your company now," Steinberger said [2].

While the demonstration focused on the security benefits of the containerized environment, other industry perspectives remain divided. Reporting from PCWorld suggests that OpenClaw AI continues to pose significant security risks despite the capabilities shown during the event [3].

Microsoft Build serves as the primary venue for the company to showcase developer tools. The integration of OpenClaw suggests a broader strategy to attract AI developers by providing specialized infrastructure for running high-risk agents securely on Windows devices [1].

“"You can totally run OpenClaw inside your company now."”

The move to integrate OpenClaw via Execution Containers represents an attempt to bridge the gap between experimental AI agency and enterprise security standards. If successful, this architecture could allow companies to deploy autonomous AI tools that can execute code and manage files without granting the software full access to the corporate network.