Snowflake, a global AI data cloud company, said on Tuesday it will unveil various AI security technologies to build a foundation for trusted agent interoperability.

Cortex AI Gateway, set to be released as a public preview, will help companies strengthen AI agent security and provide centralized visibility and control over AI usage costs.

It will support control, orchestration and management of third-party AI agents developed on external platforms such as Claude Code and Cursor, in addition to Snowflake CoWork and Snowflake CoCo.

Based on technology from Natoma, which Snowflake acquired in May, Cortex AI Gateway directly integrates Natoma Enterprise MCP platform capabilities into the Snowflake platform to support secure, enterprise-grade agent interoperability.

Mayank Upadhyay (마얀크 우파다이), Snowflake's chief security and trust officer (CSTO), said, "The center of enterprise AI is moving from data interoperability to agent interoperability, which is realized when it is built on trust in how agents on different platforms access data, call tools and support performing tasks on behalf of users, and security must be core to this transition." He said, "Snowflake provides the visibility, governance and control needed to implement AI interoperability securely in operating environments. The future of the agentic enterprise cannot be realized with only closed agent ecosystems. Snowflake is a trusted control platform that enables companies to carry out work securely."