Snowflake Horizon Catalog Unifies Governance, Context, and Security Across the Enterprise to Power Trusted AI

*This press release is a summary translation of the content announced by Snowflake Inc. on June 2.*

New innovations across the Snowflake Horizon Catalog unify AI governance, context, and security, providing a trusted foundation for enterprise AI across data, tools, and agents.

Designed for the Age of Agents:

New enhancements across the Snowflake Horizon Catalog enable all people, tools, and agents to share a common, trusted business context through richer semantic views and built-in governance and security controls. Customers like BlackRock are leveraging the new Horizon Context to operationalize AI based on a common definition of 'the single source of truth'.

Proactive Security for the AI Era:

Acxiom, NewDay, and Thomson Reuters are collaborating with Snowflake to validate how Snowflake's new AI security features can enhance security, governance, and control in enterprise AI and agentic systems.

Simple Scalability:

By seamlessly integrating with the Snowflake Horizon Catalog, Adaptive Compute enables the large-scale deployment of AI apps and agents, balancing superior performance with operational simplicity.

Snowflake (NYSE: SNOW), the AI Data Cloud company, today announced at Snowflake Summit 26 new innovations to the Snowflake Horizon Catalog that redefine how enterprises govern, contextualize, and secure AI.

As organizations move from AI experimentation to autonomously operating systems at enterprise scale, they need AI that operates on trusted business context while maintaining security, governance, and compliance. The Snowflake Horizon Catalog serves as a common AI catalog for enterprise data. New features like Horizon Context ensure all users, tools, and AI agents operate based on the same trusted business context. Combined with new security innovations focused on governing and securing AI agents, the Snowflake Horizon Catalog becomes an integrated foundation for trusted AI. Based on the governance, security, and controls provided by the Horizon Catalog, Adaptive Compute automatically optimizes compute and software resources in real-time on behalf of customers. This delivers fast, efficient AI and application performance at enterprise scale without the need for manual tuning or infrastructure management.

Christian Kleinerman, Executive Vice President of Product at Snowflake, said:

"In an era where intelligence is becoming autonomous, trust is not an afterthought—it is the foundation. Organizations need AI that operates on trusted business context with governance and security built in from the start. With this new evolution of the Snowflake Horizon Catalog, we provide all agents, apps, and people with the trusted context and security controls needed to move AI from experimentation to production."

Building a Common Understanding of Business Data for AI

As AI agents make more autonomous decisions, even minor inconsistencies in data can lead to significant errors. This is why many AI projects stall, failing to move from proof-of-concept (PoC) to production. Traditional semantic layers are decoupled from the data itself, making it difficult to maintain consistency in definitions and governance across multiple systems. For example, if an AI agent suggests a price increase based on revenue data, but the definition and calculation method for revenue differ across systems, the suggestion could lead to incorrect decisions.

Horizon Context solves this by providing a context layer for AI and BI, unifying the meaning of data across the company to ensure the reliability of AI-driven decisions. Companies like BlackRock are using Horizon Context to ensure their AI operates on a common definition of 'the single source of truth'. Horizon Context achieves this through the following capabilities:

Collecting Trusted Business Context:

Currently, business logic is fragmented across SQL, BI dashboards, and agents, undermining the reliability of AI-driven insights. Horizon Context integrates business context across an organization's data assets, including databases, data lakes, and BI tools, ensuring all tools, teams, and AI agents reference the same trusted context. This allows teams to quickly identify, organize, and confidently use the right data.

Automatically Maintaining Business Context:

When LLMs operate based on business logic, they stop guessing and start reasoning accurately. Horizon Context allows data analysts to integrate business context—how data is used, trusted, and updated—for a richer, more connected data view. With Semantic Studio, common business logic can be defined without SQL expertise, and Semantic View Autopilot automatically generates and improves semantic views that continuously maintain this context. For data shared across organizations, such as datasets from the Snowflake Marketplace, semantic views and data agents can be automatically created, reflecting trusted business context in all AI and analytics workflows. By giving AI access to consistent, governed business definitions, asking business questions to Snowflake CoCo, a coding agent for faster development, yields trusted answers even when the data resides outside Snowflake.

Extending Trusted Context to Every Environment:

Snowflake is extending access to trusted business context across the enterprise, enabling more reliable