The International Telecommunication Union is stepping into agent identity. On July 9, 2026, the ITU announced its Focus Group on Trust and Identity for Humans and Agentic AI, a new effort to build international frameworks for verifying who or what an autonomous agent actually is before it acts on a network. The group's mandate is broad: it's expected to study common terminology, identity and trust architectures, agent discovery, credential interoperability, lifecycle models, security criteria, and benchmarks, laying groundwork for a formal standards roadmap. This follows a pattern that's been building all year. Singapore's IMDA published the first comprehensive agentic-AI governance framework in January, requiring a verifiable digital identity and an audit trail for every agent action. NIST's Center for AI Standards and Innovation launched its own AI Agent Standards Initiative in February, explicitly framing agents as service accounts without dedicated identity or accountability controls attached. What's different about the ITU move is scope: it's the first attempt to pull this into a UN-level, multilateral standards body rather than leaving it to national regulators or individual vendors. The practical stakes are less about whether agents need credentials — that part is largely settled — and more about who controls issuance. A workable system needs to support multiple issuers, pseudonymous identities, selective disclosure of attributes, short-lived permissions, and transparent revocation, so that no single platform or government becomes a gatekeeper for which agents are allowed to transact. Get that wrong and an open standard curdles into a closed admission system, where a handful of identity vendors or major platforms decide which agents count as legitimate simply by choosing whose credentials they'll accept. This is also where the commercial layer is racing ahead of the standards layer. Microsoft's Entra Agent ID went generally available in May, giving each agent its own identity inside the Microsoft ecosystem, while CrowdStrike, Okta, and CyberArk have all shipped competing agent-identity products in the last two quarters. None of these interoperate by default. The ITU focus group won't resolve that fragmentation on its own, but it's the clearest signal yet that agent identity is moving from an enterprise IT problem to something international bodies think needs a shared, cross-border floor before agents are transacting across jurisdictions at scale.