Google Cloud Wants Real-Time Agent Trust Scores. Okta Launches Agent Identity Platform April 30. The Race Is On.

Two announcements in 24 hours that tell you exactly where agent identity is heading. Google Cloud: Trust Is Not Static Google Cloud's new paper on securing agentic AI at the edge makes a statement that should become a design principle: Since trust should not be seen with a static perspective, we envision a system where an agent's "trust score" is monitored and assessed in real-time. Their architecture: Hardware root of trust via TPM and secure elements — cryptographically validate the agent before it even boots Real-time behavioral monitoring — if a GDPR-certified agent tries to export raw video instead of anonymized insights, revoke credentials instantly Identity anchored in execution environment, not just registration artifacts This is Google saying: OAuth tokens and API keys are not enough. You need continuous trust assessment. Okta: Agent Identity Goes Enterprise Okta for AI Agents launches April 30, 2026. It's a platform for discovering, registering, and managing AI agents — inclu