Enterprises running AI agents across data, apps and platforms can now secure and monitor that activity through Cortex AI Gateway, a new control layer from Snowflake.
The gateway lets organisations govern how first-party and third-party agents access models, tools, MCP servers and enterprise systems. It also intelligently routes agent requests and tracks token usage to help prevent runaway AI costs. Snowflake said the tool addresses two of the biggest barriers to enterprise AI adoption, securing agents and controlling AI consumption costs.
The launch reflects a shift the company describes as moving from data interoperability to agent interoperability. As AI agents increasingly collaborate across enterprise data, apps and platforms, they introduce security and governance risks that traditional architectures were not designed to handle. Many organisations also lack a central way to see and govern how much AI they are consuming.
Snowflake said Cortex AI Gateway governs both agents built within its platform, such as Snowflake CoWork and CoCo, and external agents developed on tools including Claude Code and Cursor. Teams gain a single point of visibility and control across AI access, agent activity and token usage.
The gateway builds on Snowflake's May 2026 acquisition of Natoma, bringing that company's enterprise Model Context Protocol platform directly into Snowflake for secure agent interoperability. MCP is the emerging open standard that lets AI agents connect to external data sources and tools, and its rapid adoption is a large part of why agent governance has become pressing.
Alongside the gateway, Snowflake announced secure third-party agent access integrations with Aembit, 1Password, Linx Security, Okta, SailPoint and Saviynt. The company said this set a new standard for how enterprises govern and audit AI agents across the security ecosystem.
"Enterprise AI is moving from data interoperability to agent interoperability, and security has to be at the center of that shift," said Mayank Upadhyay, chief security and trust officer at Snowflake.
Upadhyay said agent interoperability only works when enterprises can trust how agents from different platforms access data, invoke tools and take action on behalf of users. He said Snowflake aimed to provide the visibility, governance and control needed to make that interoperability secure for production AI.