What happened
Snowflake announced Cortex AI Gateway and production-ready AI security features at Black Hat 2026, integrating a centralized MCP (Model Context Protocol) gateway called Natoma to govern how AI agents access models, data and enterprise tools across first-party and third-party ecosystems including Amazon Bedrock, Azure AI Foundry, ChatGPT, and Claude Code.
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 97% of organizations are committed to implementing AI but 57% 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, and Snowflake's approach embeds security controls directly into the data and control planes rather than relying on legacy monitoring tools.
What to watch
Cortex AI Gateway features are rolling out in stages—Wide Model Catalog, Access Governance and Sprawl Control, Agent Action Auditability, AI Cost Control, and Intelligent Model Routing are in private preview; Agent Identity and Restricted Session Scope are moving to general availability; Native AI Security Posture Management is now GA; and Advanced Data Exfiltration Prevention and Client-side CoCo CLI VM Sandbox are in preview.
Summaries like this, in your inbox every morning.
The surge in AI security concerns reflects a structural gap between enterprise ambition and operational readiness. The Linux Foundation data shows security worry nearly tripling in two years, while three-quarters of organizations struggle with capacity constraints in this area. Snowflake's announcement directly addresses this: the company is positioning security not as a bolted-on compliance layer but as a core function embedded in the data and control planes where agents operate.
Autonomous agents create a new attack surface because they combine data access, system execution, and data movement in a single profile—a qualitatively different risk from traditional user-centric access control. Decentralized agent deployment across multiple LLM providers and tools (ChatGPT, Claude, Bedrock, etc.) has created visibility and sprawl problems; teams lack a unified view of which agents can do what, at what cost, and with what audit trail. Snowflake's Cortex AI Gateway solves this by placing a centralized governance layer—based on the MCP standard and the Natoma gateway—between agents and enterprise resources.
The release strategy itself is telling: core identity and governance features (Agent Identity, Restricted Session Scope) are moving to GA or GA-soon, signaling production readiness; cost control and advanced observability remain in private preview, indicating these are evolving. This staged approach lets enterprises adopt foundational controls immediately while Snowflake refines the more sophisticated capabilities.
Pick your industry and the AI tools you use, and get news related to your work every day.
Free · 30 seconds with Google · unsubscribe anytimeWhat is AIToday? →
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.
