Snowflake Summit 2025 Highlights: Building the Future of AI and Apps

A strong data foundation has never been more important as more enterprises look to generate value with AI and applications across the full data lifecycle and make those benefits available to workers throughout their organization. Snowflake is committed to helping companies realize that goal, so we’re setting out to reimagine data engineering, analytics, app development and collaboration — and enterprise AI, of course. At Summit 2025, we’re showcasing a platform that’s now even easier to use, that quickly connects you to the data, people, apps and agentic AI products that make your business thrive, and delivers unwavering trust that your sensitive data remains protected.

This blog spotlights some of the top features and enhancements announced at Summit 2025 — but there’s much more to explore. For deep dives into the latest news and updates across the Snowflake ecosystem, check out our other Summit announcement blogs.

Bridging the gap between data and action with agentic AI

Together, today’s AI and ML feature launches form a unified AI foundation that simplifies development, scales reliably and preserves trust within Snowflake's governed environment. Fully managed within Snowflake’s secure perimeter, these capabilities enable business users and data scientists to turn structured and unstructured data into actionable insights, without complex tooling or infrastructure.

• Snowflake Intelligence (public preview soon) offers a new agentic experience (accessible through ai.snowflake.com) that gives business users the ability to securely converse with data using natural language, helping them to not just find answers but ask complex, detailed questions and take action, all from a single interface and without writing code. It brings together insights from structured and unstructured data and — because it runs within the Snowflake perimeter — inherits the benefits of Snowflake’s built-in governance and data privacy features. Snowflake Intelligence is powered by large language models from Anthropic and OpenAI, running inside the Snowflake perimeter, and Cortex Agents (public preview) under the hood. 

• Data Science Agent (private preview soon) is an agentic companion that can boost productivity for data scientists by automating every step of ML workflows with natural language, including data prep, feature engineering and training. 

• Cortex AISQL (public preview) brings multimodal data processing using AI into familiar SQL syntax. It makes complex AI workflows accessible by enabling teams to analyze documents, images and other unstructured data formats using SQL. It complements Snowflake's comprehensive unstructured data insights offering, including enhanced Document AI with schema-aware table extractions (public preview) to pull structured tables from complex PDFs and enhanced retrieval via Cortex Search.

• AI observability in Snowflake Cortex AI (generally available soon) enables no-code monitoring of generative AI apps. Snowflake also provides access to LLMs from OpenAI through Microsoft Azure OpenAI Service, Anthropic, Meta, Mistral and other leading providers, all within Snowflake’s secure perimeter. 

Learn more about these and other AI/ML announcements here.

Empowering data engineers by unlocking data interoperability

When data engineers have to wrangle data pipelines into place and work around inflexible systems, it’s going to take longer to get data from where it’s created to where it can be used — and that can stifle business success. Snowflake is committed to changing that by giving data engineers the tools and platforms to navigate the modern data landscape with confidence. Here are a few of the innovations you’ll see at Summit:

• Introducing Snowflake Openflow: An open, extensible, managed multimodal data ingestion service designed to simplify data movement and integration, Snowflake Openflow (generally available on AWS) is powered by Apache NiFi and aims to eliminate data silos and manual labor in data ingestion. Openflow revolutionizes data movement within Snowflake to enable seamless ETL processing for AI. All data integration is in one platform, and hundreds of connectors and processors simplify data integration with a variety of data sources and strategic partners. For example, Snowflake is partnering with Oracle on a solution for replicating change data capture from Oracle databases to Snowflake.

• dbt Projects on Snowflake: With this native option (public preview coming soon), data teams can build, run and monitor dbt Projects directly in Snowsight UI, accelerating the development lifecycle of data pipelines. dbt Projects is available in Snowflake Workspaces, a new native file-based development environment with features like inline AI Copilot code assistance and native Git integration.

• Enhanced Apache Iceberg^TM table support: Rethink how you build open, connected and governed data lakehouses. Integrate any Iceberg REST-compatible catalog to Snowflake, including Snowflake Open Catalog, to securely read from and write to any Iceberg table with Catalog Linked Databases (public preview soon). Build declarative pipelines with Dynamic Iceberg tables, and unlock value from semi-structured data with VARIANT data types (private preview). Optimize writes and read more of your Iceberg ecosystem with Merge on Read (private preview).

• Modern DevOps enhancements: Alongside the release of Workplaces for dbt and SQL filies, you’ll find support for custom Git URLs, the release of our Terraform provider to GA and support for Python 3.9 in Snowflake Notebooks.

Get the complete rollup of data engineering news from Summit in this blog.

Boosting AI-powered analytics and data migration

Harnessing AI can mean grappling with complex infrastructure, high costs, arduous data warehouse migrations and overloaded data teams. To help enterprises overcome those barriers, Snowflake is announcing a series of advancements focused on making analytics faster, easier to use and more intelligent — and making it easier for organizations to migrate from legacy systems and take full advantage of these capabilities: 

• Snowflake Semantic Views (public preview) help bridge the gap between raw data and business understanding by allowing you to define and store business metrics and entity relationships within Snowflake. This helps AI assistants and BI tools derive more accurate and consistent results.

• Standard Warehouse – Generation 2 (Gen2) is an updated version of Snowflake’s current Standard Warehouse that has upgraded hardware and additional performance enhancements such as improvements to Delete, Update and Merge operations and to speed up table scan operations. Over the past 12 months ending May 2, 2025, Snowflake has delivered 2.1x faster performance for core analytics workloads on Snowflake tables through Gen2 (see more details here).

• SnowConvert AI is a free, automated solution designed to help accelerate data warehouse, BI and ETL migrations from legacy platforms. It intelligently analyzes your existing code, automating code conversion and data validation while streamlining the entire migration process, lowering the risk associated with large-scale data migrations.

Read this blog to learn more about these announcements and others that will help data analysts and architects transform raw data into actionable insights with ease and speed.

Empowering innovation with seamless collaboration 

Shifting from AI experimentation to delivering real business value lies in how well you can leverage not just your own data, but that of external parties in your ecosystem — how easily you can share data, connect to third-party sources, and build and collaborate with them. With that in mind, Snowflake is aiming to make it easier to connect to and collaborate with AI-ready internal and external sources, extend agentic workflows, and build and deploy AI-driven apps while protecting private data. Here’s a look at some of the flavors of agentic products launching on Snowflake Marketplace:

• Cortex Knowledge Extensions (generally available soon): Enrich apps with real-time content from publishers like USA TODAY, The Associated Press, Packt, Stack Overflow and CB Insights while protecting intellectual property and supporting proper attribution.

• Sharing of Semantic Models (private preview): Allows users to integrate and share AI-ready internal data sets or third-party structured data from internal teams and third-party providers, and use natural language to “talk to data” from a variety of providers.

  Agentic Snowflake Native Apps on Snowflake Marketplace: A new way for providers to develop and offer Snowflake Native Apps that reference Cortex Agent APIs to automate tasks and interact with data, these interoperable agentic products can give data engineers, data scientists and others a fast track to deploying and creating value from agentic AI.

• Snowflake Native App Framework enhancements for security and interoperability: New features focused on application versioning, permissioning, app observability and compliance badging make it easy to build security into apps from the ground up.

Get more details and see how Snowflake Native Apps, Secure Data Sharing and Snowflake collaboration capabilities can benefit your company in this blog.

Redefining expectations for a modern data platform

In the age of AI, a robust platform is a must-have. But to deliver the most value, the platform must also dynamically adjust resources based on workload demands, deliver high performance without driving up costs, protect and govern sensitive data, and be easy for users across the enterprise (not just analysts and engineers) to use. Snowflake’s platform advancements target these common challenges by raising the bar for modern data infrastructure:

• Snowflake Adaptive Compute (private preview): Make warehouses even easier to use by automatically scaling resources and intelligently routing queries to reduce customers’ platform management burden. Warehouses created using Adaptive Compute, known as Adaptive Warehouses, accelerate performance for users without driving up costs. With Adaptive Warehouses making infrastructure virtually invisible to users, teams can focus more time on strategic projects that drive revenue and less on warehouse maintenance.

• Snowflake Horizon Catalog interoperability expansion: Along with support for governing and discovering data in Iceberg tables, Horizon Catalog will provide external data discovery (private preview soon) for external data in relational databases, dashboards, semantic models and more.

• Copilot for Horizon Catalog (private preview soon): Use natural language to perform governance, security and metadata discovery tasks in Snowsight, simplifying the process of getting critical information about data sets.

• Security: New Snowflake Trust Center extensions (generally available soon), new MFA methods and account security updates, password protection and other enhanced security features aim to strengthen account protection.

• Out-of-the-box observability: Improvements to Snowflake Trail help you better understand your infrastructure and applications with Snowpark Container Services, analyze your pipelines with telemetry support for Snowflake Openflow, and debug and optimize your generative AI agents and apps (generally available soon).

Learn more about these enhancements and other updates to business continuity capabilities, Unistore and more in our platform-focused blog.

Forward Looking Statements

This article contains forward-looking statements, including about our future product offerings, and are not commitments to deliver any product offerings. Actual results and offerings may differ and are subject to known and unknown risk and uncertainties. See our latest 10-Q for more information.