I just got back from this year’s Snowflake Summit, and the energy was as high as ever. Keynotes were packed, booths buzzing, and the Porsche giveaway had people genuinely wondering if they might drive home with more than swag.
Unsurprisingly, “AI” was the word of the week, but the real story was how large enterprises are finally putting down strong data foundations and maturing their cloud strategies. (As a San Francisco local, I was especially excited about the announcement of the new Silicon Valley AI Hub (SVAI) in Menlo Park. It feels like a major step forward for live collaboration in the Bay Area.)
Here’s a recap of what stood out most from the event and the latest announcements.
The AI era of Snowflake: Now very real
Snowflake has gone beyond buzzwords with some serious investment in AI tooling:
- AI SQL Functions: With commands like AI_Filter() and AI_Aggregate(), you can use natural language to query multimodal data. This isn’t just SQL generation — it’s AI inside your SQL.
- Cortex knowledge extensions: Think live context from external data sources like the New York Times, SharePoint, and Stack Overflow. Combine this with Cortex Agents (coming soon) and you’ve got a framework for AI-powered insight and action.
- Marketplace-ready AI tools: Soon, you’ll be able to publish and deploy Cortex Agents that can answer questions, trigger actions in tools like Slack or Teams, and surface search results, all inside the Snowflake ecosystem.
This isn’t future-state — it’s happening right now. And it marks a serious shift in how we’ll build with data.
Bringing the work to the data (literally)
My favorite updates were about dbt:
- Run dbt projects in Snowflake: Within a new Snowflake Workspace (connected to a git repo), users can compile, run, and test projects directly.
- The dbt Fusion Engine: A complete overhaul of the backend that powers dbt, bringing lightning fast compilation, robust project parsing, and more options for interfacing with AI.
Snowflake is investing in all the other workloads too:
- Snowflake Postgres (from their CrunchyData acquisition): Brings transactional capabilities to Snowflake in ways that Hybrid Tables never could.
- Workspaces: A fully integrated IDE in Snowflake tied to your remote git repo.
- Semantic views: A new object type for defining reusable metrics via “Semantic SQL” which will enable agentic workflows.
Platform enhancements you shouldn’t miss
This year’s list of new features was long, but here are a few that made me pause:
- Gen2 Warehouses: Claim 2x performance with the same credit usage
- Openflow: A managed ingestion service built on Apache NiFi; simplifies real-time CDC from multiple sources — and is maturing fast
- Tag-Based Budgets and Privacy Controls (Projection, Aggregation, and Differential Privacy Policies): Give enterprise teams more governance power with less friction
Also worth noting: Snowpipe Streaming just got faster and cheaper, and Snowpipe pricing will soon shift to being based on GB size, not file count.
Simplicity over complexity
While talking to customers at the event, I heard some consistent themes:
- Organizations are prioritizing data products over datasets
- They’re pursuing architectural simplicity
- And deliberately building foundational models that set the stage for AI
Snowflake is meeting these goals by abstracting complexity across the board. Whether it’s Adaptive Compute Warehouses that auto-tune performance, or AI-assisted migrations with SnowConvert, the platform is leaning into user experience without sacrificing control.
Looking ahead
If one thing is clear from this year’s Summit, it’s that Snowflake’s north star remains unchanged: bring the work to the data. But the way they’re executing that vision has evolved.
They’re enabling transactional workloads, AI development, and live data engineering — all in one place. And they’re doing it with simplicity, performance, and developer tooling that feels thoughtfully built.
An experienced Snowflake partner for the AI-powered future
We’re excited at Atrium to bring these innovations into the work we do with clients. If you’re thinking about how to take advantage of Snowflake’s new capabilities, let’s talk.
As an Elite Snowflake Services Partner, Atrium helps you get more out of your investment in the AI Data Cloud, from migration and implementation to managed services and data science integration.
