What happened
At Snowflake Summit 2026, SiliconANGLE's theCUBE interviewed 1Password CTO Nancy Wang and Snowflake senior director of product management Artin Avanes about securing AI agents. Wang said, "The conversation has really changed from how do we secure systems to now how do we govern actors," adding that agents make the problem "truly multidimensional." Avanes said, "There is no world where you can just put an agent out or run your agentic tool directly against your data without those controls." Wang also framed the shift as one where "identity security now meets data security." SiliconANGLE discloses that theCUBE is a paid media partner for Snowflake Summit.
Background
Per the 1Password blog dated April 30, 2026, 1Password expanded a partnership with Perplexity to integrate secure access for Perplexity Computer, with Perplexity CBO Dmitry Shevelenko quoted that "secure and seamless access has to be built into the experience from the start." Per SiliconANGLE's April 22, 2025 report, 1Password earlier launched Agentic AI Security features and extended its Extended Access Management (XAM) platform to govern human and nonhuman access, including an SDK for managing secrets in agentic workflows.
Why it matters
Editorial analysis: Treating agents as first-class actors implies runtime secrets handling, least-privilege enforcement, dynamic policy, and audit trails that cover machine identities, not only human users. Avanes tied accuracy and trust to Snowflake's Horizon Context, which taps metadata outside the warehouse to ground agentic workflows, reflecting a broader industry move to couple data catalogs and semantic layers with agent execution.
What to watch
Editorial analysis: Watch how secret stores, IAM and policy engines, and data platforms integrate to issue ephemeral or credential-less access for agents, and whether compliance and telemetry tooling extends to nonhuman actors. Also track emerging standards for agent identity, attribution, and runtime policy enforcement from major cloud and security vendors.
Key Points
- 11Password and Snowflake executives argue AI agents form a new class of actors requiring runtime governance, identity controls, and auditability, not just perimeter security.
- 2Snowflake tied the discussion to its Horizon Context metadata and governance capabilities, positioning semantic context as a way to make agent workflows more accurate and controllable.
- 3The segment is vendor-framed theCUBE conference coverage (a paid media partner), useful as a signal of enterprise priorities rather than independent evaluation.
Scoring Rationale
Securing AI agent access and runtime secrets is a real and timely operational concern for teams moving agents into production, and Snowflake did surface concrete governance work (agent identity, Horizon Context). But the event itself is a vendor-framed theCUBE interview at a sponsored event, not a product launch or independent evaluation, so it lands as solid-but-promotional rather than notable industry news.
Practice interview problems based on real data
1,625 SQL & Python problems across 15 industry datasets — the exact type of data you work with.
Try 250 free problems
