Highlights from Snowflake Summit 2025

Making Data More Intelligent, Trusted and Conversational

by Kathryn Nguyen, Data Consultant

Earlier this month, I had the privilege of attending Snowflake Summit 2025, and the energy around data innovation was unmistakable!

From AI-native features to enhanced governance and developer workflows, Snowflake is leaning in heavily on a future that makes data more accessible, intelligent, and secure without compromising performance or control.

Here are some of my top takeaways from the Summit that I believe will reshape the way businesses approach data in the next 12 months.

AI & Agentic Intelligence: Making Data Conversational

One of the most exciting announcements was the unveiling of Snowflake Intelligence—a new natural language interface available at ai.snowflake.com.

This interface lets business users query both structured and unstructured data, launch workflows, and act on insights—all securely within the Snowflake environment.

It’s powered by large language models from OpenAI and Anthropic, orchestrated through Cortex Agents, which handle everything from search to task automation.

Another standout is Cortex AISQL—which brings AI capabilities like AI_FILTER and AI_CLASSIFY directly into SQL.

That means analysts can now process multimodal data such as text, images, and audio without leaving their SQL environment.

This dramatically reduces friction between insights and execution.

We’re also watching Data Science Agent closely—an upcoming tool to automate ML workflows using natural language, simplifying the end-to-end process from data prep to model deployment.

What this means for our clients

We’re entering an era where data conversations are no longer confined to dashboards and analysts. These tools will make insights more immediate, contextual, and integrated across every part of the business.

Governance & Metadata: Trust at Scale

Snowflake continues to raise the bar in governance with features designed to ensure trust and compliance at scale:

• Horizon Catalog & Copilot: Horizon now includes automatic sensitive data tagging and end-to-end lineage tracking, supported by an AI assistant that helps users understand and manage data assets in context.
• Semantic Views: This is a game-changer for BI and AI teams as it lets you define consistent metrics and business logic across your organisation. It’s particularly helpful when aligning AI models to business definitions.
• Unified Policy Layer: Now extended across SQL analytics, transactional processing, and AI inference, this feature ensures governance policies are enforced consistently, reducing risk in hybrid data architectures.

Why its important

For enterprise clients juggling compliance, cost control and cross-team visibility, these updates will simplify auditability and governance without slowing down innovation.

Platform Enhancements: Speed, Simplicity, Scale

Performance improvements and developer experience upgrades were a strong focus this year:

• Gen2 Warehouses: Delivering 2.1x faster analytics performance with no extra tuning. More speed, less complexity.
• Adaptive Compute (Private Preview): Dynamically scales compute resources based on workload—ensuring cost efficiency and responsiveness.
• OpenFlow: Built on Apache NiFi, this managed data ingestion service simplifies real-time data integration from diverse sources and formats.
• dbt Integration: You can now run dbt projects directly inside Snowflake, streamlining data transformation and analytics engineering.

What excites us

OpenFlow opens the door for seamless ingestion of live and external data sources, making real-time analytics more achievable. And for Snowflake users already invested in dbt, this native integration means faster, more governed workflows.

What We’re Watching Next

Among all the innovations, cost-based anomaly detection using AI stood out.

The ability to proactively monitor and trigger alerts on usage spikes could significantly reduce unexpected cloud bills—a growing concern for many data teams.

To get the most out of these alerts, we’ll need to make sure they’re set up to trigger at the right time and for the right reasons.

At Data Army, we’re exploring how to embed these new AI features into our own Snowflake-powered accelerators and governance frameworks, ensuring clients gain faster insights without increasing complexity.

Final Thoughts

Snowflake is no longer just a data warehouse. It’s rapidly becoming an intelligent, policy-aware data operating system—bridging the gap between raw data, business logic, and human decisions.

For our enterprise clients, this means more control, better performance, and a stronger foundation for AI-driven decision-making at scale.