Snowflake Summit 2025: Empowering Builders with AI
Unpacking the innovations transforming data into intelligent action
At Snowflake Summit 2025 Builders Keynote, the spotlight was on empowering developers, data engineers, and analysts with cutting-edge AI tools. Key announcements included:
Here’s a summary of the key announcements and highlights from the Snowflake Builder Keynote:
🔑 Major Themes
AI is central to Snowflake’s future and transforming the software development lifecycle.
Focus on empowering builders—data analysts, data engineers, AI/ML engineers—with new tools, frameworks, and community resources.
Emphasis on agentic systems, real-time and unstructured data support, and open standards like Apache Iceberg.
🚀 Key Product Announcements
1. Snowflake Intelligence
New UI layer enabling organizations to use AI agents across business functions.
Built on Cortex AI platform for scalable, trustworthy agentic systems.
Includes semantic modeling for accurate, controlled AI outputs.
2. Cortex Search API
Powered by Snowflake Arctic Embed model (open-sourced on Hugging Face).
Optimized for semantic search across structured and unstructured data (docs, audio, images).
3. AI Agents & Agentic Systems
Showcase of real-world implementation by Whoop using Cortex tools for internal data search and analytics agents.
Agents can choose tools, interact with structured and unstructured data, and follow semantic guardrails for accuracy.
4. Streamlit in Snowflake
Now generally available on AWS and Azure.
Includes major open-source features: file upload, camera/audio input, and custom components (public preview).
Formation of Streamlit Steering Committee to guide future development.
5. Snowflake Notebooks
Generally available on AWS, Azure, and GCP.
Now supports custom runtime environments (e.g., containers) for advanced ML workloads.
Supports private connectivity across clouds.
6. Snowflake Workspaces
New one-stop UI for managing files, charts, and data projects.
Integrated AI copilots, improved SQL support, and dbt integration.
7. Open Standards & Apache Iceberg
Expanded support for externally managed Iceberg tables and Apache Polaris catalog.
Support for Iceberg V3 features (e.g., geometry, variant data types).
Snowflake committed to interoperability and open data architecture.
📡 Real-Time and Unstructured Data
Document AI enables extracting data from unstructured formats into structured Snowflake tables.
OpenFlow supports real-time streaming data pipelines for AI applications.
🧠 LLM and Inference Enhancements
Cortex supports top models (Claude 4, Mistral, Meta, Snowflake Arctic).
Speculative decoding and suffix caching for faster inference.
Plug-and-play OpenAI SDK compatibility via Snowflake endpoints.
🧪 Community & Ecosystem
Silicon Valley AI Hub (Coming Fall 2025)
28,000 sq. ft. co-working and collaboration space for AI developers, startups, and Snowflake partners.
Git & DevOps Enhancements
Support for custom Git repo URLs, access tokens, and private OAuth workflows.
🎷Cultural & Closing Notes
Analogy of developers as a jazz band—diverse talents, shared purpose.
Call to action: Build agentic systems, embrace evolving roles, and participate in shaping AI’s future in the Snowflake ecosystem.
Snowflake Data Superhero and Chapter Lead for the Dutch Snowflake ❄️ User Group. Consulting Partner at Bravinci. Online also known as; DaAnalytics
