What is Model Context Protocol (MCP) and why does Snowflake's gateway govern it?

Model Context Protocol is a standard that enterprises use to connect LLMs to databases, internal tools and SaaS environments. Decentralized MCP adoption creates unmanaged sprawl, fractured user experiences and security vulnerabilities such as unvetted servers, tool hijacking and data exfiltration. Cortex AI Gateway, by integrating Natoma (a centralized MCP gateway), enforces identity, policy and audit at the tool-call level to give enterprises visibility and control.

Which AI models and platforms does Cortex AI Gateway support?

Cortex AI Gateway governs how agents access both first-party tools (Snowflake CoCo and CoWork) and third-party ecosystems including Amazon Bedrock, Azure AI Foundry, ChatGPT, Claude Code, Cursor, custom LangChain or LlamaIndex apps and others. A Wide Model Catalog in private preview will bring GPT, Gemini, Claude, Grok, Mistral and GLM under one roof, allowing enterprises to run them in their geography to keep data within the required region.

What new security features are now available, and which are still in preview?

Features moving to general availability (GA) or soon to GA include Agent Identity (visibility into agent activity and audit tracking), Restricted Session Scope (limiting agent sessions to required permissions), Native AI Security Posture Management (integrated dashboard in Snowflake Trust Center for proactive risk scanning), and Ransomware Protection via Multi-Party Approval (requiring two or more authorizations before destructive changes). Features in private preview include Wide Model Catalog, Access Governance and Sprawl Control, Observability and Tracing, AI Cost Control, Intelligent Model Routing, Context-Aware Access Policies, Advanced Data Exfiltration Prevention, and Client-side CoCo CLI VM Sandbox.

AI Safety & AlignmentAI Business & IndustrySnowflake AI BlogAug 3, 2026Aug 3, 2026, 13:00 JST

Snowflake launches AI Gateway to govern autonomous agent sprawl

Send on LINE

3 Key Points

What happened
Snowflake announced Cortex AI Gateway at Black Hat 2026, a centralized control layer that enforces identity, policy and audit over autonomous agents accessing models, data and enterprise tools via Model Context Protocol (MCP) servers. The company also moved several AI security features to general availability (GA) or public preview, including Agent Identity, Restricted Session Scope, Native AI Security Posture Management, Ransomware Protection via Multi-Party Approval, and Data Exfiltration Prevention.
Why it matters
AI security concerns have surged from 17% in 2024 to 48% in 2026 according to The Linux Foundation's 2026 State of Tech Talent Report, while 57% of organizations face a significant capacity gap in security and risk management. Autonomous agents are expanding the enterprise attack surface by combining data access, system execution and data movement; Cortex AI Gateway addresses the hard problem of fragmented access, lack of visibility into agent actions and spiraling AI costs by providing centralized governance across first-party tools (Snowflake CoCo, CoWork) and third-party ecosystems (Amazon Bedrock, Azure AI Foundry, ChatGPT, Claude Code, Cursor, LangChain, LlamaIndex).
What to watch
Cortex AI Gateway offers three core capabilities in its initial release: Control (grant, restrict and audit access from one endpoint), Visibility (capture agent actions in real time for audit trails), and Cost and Performance (automatic model routing based on cost, latency and capability, with spending limits per team or workload). Several features including Wide Model Catalog, Access Governance, Observability and Tracing, AI Cost Control and Intelligent Model Routing are in private preview. Visit Snowflake booth #8206 at Black Hat USA 2026 for demonstrations.

Not sure about something? Ask the AI

Questions and answers are published on this page.

Summaries like this, in your inbox every morning.

Get news like this every morning, free

Context & Analysis

The surge in AI security concerns—from 17% in 2024 to 48% in 2026—reflects a fundamental challenge facing enterprises: while 97% of organizations are committed to implementing AI, 57% lack the security and risk management capacity to do so safely. Autonomous agents have become a flashpoint because they consolidate data access, system execution and data movement into a single profile, creating a wider attack surface than traditional application layers. Legacy point solutions and application-layer fixes alone are insufficient; security must be embedded in the data and control planes themselves.

Snowflake's announcement addresses this architectural gap by introducing governance at the agent orchestration layer. By integrating Natoma—a centralized MCP gateway—into the Snowflake ecosystem, the company is tackling the fragmented sprawl of emerging agent frameworks. Enterprises are adopting multiple autonomous agent types across different LLM providers (ChatGPT, Claude, Grok), development tools (Cursor, LangChain) and cloud platforms (Amazon Bedrock, Azure AI Foundry), each with its own access patterns and audit requirements. A centralized gateway enables teams to enforce a single set of identity, authorization and audit policies across this heterogeneous landscape, eliminating the manual burden of configuring security for each new agent type.

FAQ

What is Model Context Protocol (MCP) and why does Snowflake's gateway govern it?: Model Context Protocol is a standard that enterprises use to connect LLMs to databases, internal tools and SaaS environments. Decentralized MCP adoption creates unmanaged sprawl, fractured user experiences and security vulnerabilities such as unvetted servers, tool hijacking and data exfiltration. Cortex AI Gateway, by integrating Natoma (a centralized MCP gateway), enforces identity, policy and audit at the tool-call level to give enterprises visibility and control.
Which AI models and platforms does Cortex AI Gateway support?: Cortex AI Gateway governs how agents access both first-party tools (Snowflake CoCo and CoWork) and third-party ecosystems including Amazon Bedrock, Azure AI Foundry, ChatGPT, Claude Code, Cursor, custom LangChain or LlamaIndex apps and others. A Wide Model Catalog in private preview will bring GPT, Gemini, Claude, Grok, Mistral and GLM under one roof, allowing enterprises to run them in their geography to keep data within the required region.
What new security features are now available, and which are still in preview?: Features moving to general availability (GA) or soon to GA include Agent Identity (visibility into agent activity and audit tracking), Restricted Session Scope (limiting agent sessions to required permissions), Native AI Security Posture Management (integrated dashboard in Snowflake Trust Center for proactive risk scanning), and Ransomware Protection via Multi-Party Approval (requiring two or more authorizations before destructive changes). Features in private preview include Wide Model Catalog, Access Governance and Sprawl Control, Observability and Tracing, AI Cost Control, Intelligent Model Routing, Context-Aware Access Policies, Advanced Data Exfiltration Prevention, and Client-side CoCo CLI VM Sandbox.

Snowflake AI BlogRead Original Article

Send on LINE

AI news that matters for your work, delivered every morning.

Pick your industry and the AI tools you use, and get news related to your work every day.

Get it free

Free · 30 seconds with Google · unsubscribe anytimeWhat is AIToday? →

Ask AI

Ask AI anything about this article. The AI reads this article, earlier AIToday articles, and Wikipedia, and cites its sources. Q&As are published on this page for other readers too.