Microsoft backs agentic AI foundation

Microsoft used Open Source Summit North America in Minneapolis on May 18 to tie new Linux infrastructure releases to a broader push for open standards in AI agents. The company said it will bring Azure Linux 4.0 to public preview on Azure Virtual Machines and made Azure Container Linux generally available, with a broader rollout planned at Microsoft Build on June 2. Brendan Burns, Microsoft corporate vice president and technical fellow for Azure OSS and Cloud Native, said developers need open-source foundations that are “more secure, more predictable, and easier to build apps and agents.” The event runs May 18-20, according to the Linux Foundation. ### Why did Microsoft pair Linux releases with agent governance now? Microsoft said on May 18 that the new Linux releases are aimed at cloud-native and AI workloads, linking operating-system stability to the growing use of software agents in production systems. In a company blog post tied to the conference, Burns said open source is the foundation for AI and that scaling AI workloads requires more predictable and secure infrastructure. Microsoft had already introduced an open-source Agent Governance Toolkit on April 2 under an MIT license. The company said that toolkit provides runtime security governance for autonomous AI agents and addresses the 10 OWASP agentic AI risk categories with deterministic policy enforcement. (opensource.microsoft.com) ### What is the Agentic AI Foundation, and why is Microsoft backing it? The Linux Foundation launched the Agentic AI Foundation in December 2025 with founding contributions including Anthropic’s Model Context Protocol, Block’s goose and OpenAI’s AGENTS.md. The foundation said its role is to provide a neutral home for open standards and governance around agentic AI. (opensource.microsoft.com) Microsoft’s May 18 announcement publicly aligned the company with that effort at a moment when multiple vendors are trying to define how agents connect to tools, data and external systems. The company’s conference post framed the work as part of building open source infrastructure for apps and agents rather than as a proprietary stack. (linuxfoundation.org) ### Who joined the foundation this week? The Agentic AI Foundation said on May 18 that it added 43 organizations in the past quarter, including four new Gold Members, 27 Silver Members and 12 Associate Members. The group said total membership now stands at 180 organizations. (opensource.microsoft.com) The new Gold Members are F5, GoDaddy, Stripe and TRON, according to the foundation and the Linux Foundation. The announcement said the latest additions span financial services, infrastructure, security and public-sector organizations, which the foundation said are joining to help shape standards behind “production-grade agentic AI.” (prnewswire.com) ### What concrete standards work is already under this umbrella? The Agentic AI Foundation said in December that its initial project set included Model Context Protocol, goose and AGENTS.md. Those projects address different layers of agent interoperability, from how models connect to tools and data to how agent behavior is described in shared development workflows. (prnewswire.com) The Linux Foundation said in April that another open protocol effort, A2A, had passed 150 organizations and landed in major cloud platforms with enterprise production use in its first year. That separate milestone underscored how quickly standard-setting around agents is becoming a competitive area across the broader open-source ecosystem. ### What happens next? Microsoft said Azure Container Linux will see a broader rollout at Microsoft Build on June 2. (linuxfoundation.org) The Linux Foundation’s Open Source Summit North America continues in Minneapolis through May 20, and the Agentic AI Foundation has separately outlined a 2026 events program anchored by AGNTCon and MCPCon in North America and Europe. (opensource.microsoft.com) (linuxfoundation.org)