Snowflake Advances the Trusted Agentic Enterprise Era with Unified Monitoring and Cost Management
Tag: S-2026-07-28-snowflake-cortex-ai-gateway Type: article (vendor press release) Author(s): Snowflake Inc. (quotes: Mayank Upadhyay, Chief Security and Trust Officer) Date of source: 2026-07-28 Date ingested: 2026-08-03 Authority weight: medium — primary vendor announcement, self-interested; capability claims unverified, product largely pre-preview Raw file: /_raw_sources/S-2026-07-28-snowflake-cortex-ai-gateway.md
What it claims
Snowflake announced Cortex AI Gateway, positioned as the “connective layer for all trusted agent activity”: a centralised control plane to govern how AI agents — both Snowflake-native (CoWork, CoCo) and third-party (it names Claude Code and Cursor) — access models, tools, MCP servers (100+ supported) and enterprise systems. Claimed capabilities: centralised agent access policies, authentication and permissions; a “centralized, end-to-end record of agent activity” (what each agent did, which systems it touched, sequence of steps); unified AI cost visibility with attribution and enforced spending limits; and intelligent routing of requests to enterprise-approved models. The gateway builds on Snowflake’s May 2026 acquisition of Natoma (enterprise MCP platform). Alongside it, Snowflake announced a first wave of third-party agent access integrations with 1Password, Aembit, Linx Security, Okta, SailPoint and Saviynt, framed around task-scoped access (agents should not inherit all user permissions), agent attribution and cross-platform governability. BlackRock and Thomson Reuters are named as enterprises strengthening security with Snowflake’s zero-trust agentic approach; Meltwater comments prospectively on the gateway. Availability: the gateway is “public preview soon”; some AI-security items are GA; context-aware controls and session scoping are private preview; the security-vendor integrations are “private preview soon” (Okta planned Q4 2026).
Notable quotes
“The future of the agentic enterprise will not be built in closed agent ecosystems, and Snowflake is the trusted control plane that enables secure enterprise work.” — Mayank Upadhyay, CSTO (para. 5)
“Cortex AI Gateway provides a centralized, end-to-end record of agent activity, giving teams the visibility needed to operate agents safely at scale.” (capability bullets)
What’s speculative vs. asserted
Asserted: the announcement itself, the Natoma lineage, the named partners and customers, and the availability stages. Effectively speculative: every capability claim, since the core product is not yet in public preview — “public preview soon” and “private preview soon” are forward-looking; the release carries a forward-looking-statements disclaimer. BlackRock/Thomson Reuters are not stated to use Cortex AI Gateway specifically. Neither named customer is an EU/UK-regulated FI as framed by Snowflake; any EU/UK read-across is this vault’s inference.
Topics this feeds
Snowflake — company page created from this source. Also bears on Agentic Data Access Governance (platform-native agent-governance layer) and, as a competitive mirror, Databricks (Unity AI Gateway, June 2026).
Open questions raised
Whether the “end-to-end record of agent activity” will satisfy EU AI Act Article 12-style logging or DORA ICT-access evidence expectations (no regulation is named in the release). Actual preview/GA dates. Whether any EU/UK regulated FI adopts it, and what EU data-residency options apply.