Snowflake Inc. (NYSE: SNOW) announced new features for its Horizon Catalog platform at Snowflake Summit 26, focusing on AI governance, security, and enterprise context management.

The company introduced Horizon Context, which provides a centralized system for maintaining consistent business definitions across AI applications and analytics tools. BlackRock is using this feature to ensure AI systems operate with shared enterprise data definitions.

Snowflake also unveiled new AI security capabilities, including Agent Identity features that provide verified identities for AI agents before they can access enterprise data. The company cited a McKinsey study indicating that nearly two-thirds of organizations consider security the primary barrier to scaling AI.

Companies including Acxiom, NewDay, and Thomson Reuters are working with Snowflake to test these new AI security features for enterprise systems.

The platform now includes Semantic Studio, which allows teams to define business logic without SQL expertise, and Semantic View Autopilot, which automatically creates and maintains semantic views. The system supports the Open Semantic Interchange standard for cross-platform compatibility.

Snowflake also announced Adaptive Compute, which automatically optimizes compute and software resources for AI workloads without manual configuration. This feature is designed to work with the governance and security controls provided by Horizon Catalog.

The enhancements are part of Snowflake's broader strategy to address enterprise AI deployment challenges, particularly around data consistency and security controls for autonomous AI systems.