What happened
Snowflake announced Cortex AI Gateway, which integrates a centralized MCP (Model Context Protocol) gateway called Natoma to govern how AI agents access models, data, and enterprise tools across first-party systems (like Snowflake CoCo and CoWork) and third-party platforms (Amazon Bedrock, Azure AI Foundry, ChatGPT, Claude Code, Cursor, and others). The company also transitioned several native AI security features to general availability or public preview, including Agent Identity, Restricted Session Scope, Native AI Security Posture Management, Data Exfiltration Prevention, and Ransomware Protection via Multi-Party Approval.
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, yet 57% of organizations face a significant capacity gap in security and risk management even though 97% are committed to implementing AI. Autonomous agents are expanding the enterprise attack surface by combining data access, system execution, and data movement, but decentralized agent adoption creates unmanaged sprawl, fractured visibility, and severe security liabilities—making a centralized governance layer critical for enterprises scaling agents safely.
What to watch
Cortex AI Gateway's core features—including Access Governance and Sprawl Control, Govern Every Agent Connection (covering 100+ MCP servers), Observability and Tracing, AI Cost Control, and Intelligent Model Routing—are available in private preview. The Wide Model Catalog (supporting GPT, Gemini, Claude, Grok, Mistral, GLM and others) is also in private preview. Restricted Session Scope and Client-side CoCo CLI VM Sandbox (macOS only, currently) are coming soon or in private preview; Snowflake is demonstrating these at Black Hat USA 2026 booth #8206.
Summaries like this, in your inbox every morning.
The announcement reflects a structural shift in how enterprises must approach AI security. According to The Linux Foundation's 2026 State of Tech Talent Report cited in the article, AI security concerns have nearly tripled from 17% in 2024 to 48% in 2026—a dramatic widening of the risk perception gap. Yet 97% of organizations remain committed to implementing AI, revealing a critical disconnect: while the appetite for AI adoption is nearly universal, 57% of organizations face a significant capacity gap in security and risk management. Autonomous agents have intensified this tension by expanding the attack surface beyond traditional application layers; they now combine data access, system execution, and data movement into single profiles, multiplying the potential paths for breach or misconfiguration.
Snowflake's response is to move security "into the core data and AI infrastructure" rather than rely on patchwork application-layer fixes. The Cortex AI Gateway addresses this by introducing a centralized MCP gateway (Natoma) that enforces identity, policy, and audit at the tool-call level—the granular point where agents interact with external systems. This shift from decentralized agent sprawl to governed interoperability directly tackles the vulnerabilities the article identifies: unvetted servers, tool hijacking, data exfiltration, and the visibility blindness that comes from fragmented deployments. By extending Snowflake's existing data governance framework to agent traffic, the company is betting that enterprises will accept centralization as the necessary price of safe scale.
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.
