Snowflake Advances the Trusted Agentic Enterprise Era with Unified Monitoring and Cost Management Jul 28, 2026

Snowflake unveiled Cortex AI Gateway, a new service designed to act as a centralized control layer for AI agents, enabling enterprises to govern how agents access models, tools, and enterprise systems while maintaining visibility over AI consumption costs. The gateway supports both first-party agents built within Snowflake and third-party agents from external platforms, providing a unified foundation for secure agent interoperability. By consolidating access policies, authentication, and permissions, organizations can prevent unauthorized agent activity and reduce operational complexity. The service builds on Snowflake’s acquisition of Natoma, integrating enterprise-grade agent interoperability capabilities directly into its platform.

The company also announced new integrations with security vendors including 1Password, Aembit, Linx Security, Okta, SailPoint, and Saviynt to enable consistent governance and auditing of third-party agents across platforms. These integrations allow security teams to track agent actions, limit permissions to task-specific scopes, and enforce identity-based controls, addressing risks posed by fragmented agent ecosystems. The approach shifts away from broad user credentials toward granular, verifiable agent access, improving accountability and reducing exposure to unauthorized actions. Snowflake positions these capabilities as essential for scaling AI agents in production environments.

Snowflake emphasized advancing zero-trust security principles to support the agentic enterprise, including verified agent identities, context-aware controls, and session scoping to minimize data misuse. The company highlighted collaborations with BlackRock and Thomson Reuters as examples of enterprises leveraging these security enhancements to strengthen oversight of AI-driven workflows. These measures aim to address vulnerabilities introduced by cross-system agent collaboration, where traditional security models fall short. The focus on identity verification and task-scoped access reflects a broader shift toward production-ready AI governance.

At Meltwater, Chief Technology Officer Aditya Jami described Cortex AI Gateway as a step toward making agents more observable and production-ready while maintaining security and trust. The service provides centralized records of agent activity, enabling teams to trace actions, optimize model routing, and enforce spending limits to prevent unchecked AI costs. By offering a unified view of AI consumption, the gateway helps organizations balance flexibility with control, ensuring agents operate within defined boundaries. Snowflake frames this as foundational for building a secure, interoperable agent ecosystem across the enterprise.