4 Snowflake Summit Hot Topics

What to care about and what to leave behind as analytics engineers in 2026

Snowflake this, Snowflake that. It seems like Snowflake conversation always dominates the first week of June.

While I sadly didn’t attend Snowflake Summit this year, I still kept tabs on all the important updates coming out of the summit.

Quite frankly, there usually isn’t much that excites me. The best part about the summit is meeting the community IRL and learning about the latest and greatest tools in the space.

However, this year there were a few updates and overall themes that caught my attention.

SQLMesh has risen from the dead

If you followed along with last summer’s Data Pipeline Summer challenge, then you learned about SQLMesh and its unique value prop for data transformation. It was the one real competitor of dbt.

Then Fivetran bought them…. And then dbt Labs….

I thought SQLMesh would simply disappear forever. However, it was actually contributed to the Linux Foundation, meaning its vendor neutral. dbt then announced the addition of dbt State this week.

This was the defining factor that made SQLMesh so efficient compared to dbt. It essentially means that dbt builds only what has changed instead of building everything associated with a change.

It’s faster and reduces costs.

Apparently it’s also available for dbt Core users 👀 (which is nice to hear, but also a bit surprising).

A push towards semantics

Now that everyone is starting to see the value in adopting AI and using it to improve their workflows, we are starting to see where the real cracks are in the foundation.

First, many companies don’t have the proper data models available to take advantage of agents for analysis.

Second, even if they have data models, they don’t have clear documentation on what fields means, how to calculate certain metrics, or why certain business decisions were made. This means AI agents are inferring answers using what’s available to them.

For these agents to be effective, you need this context. Snowflake launched Semantic View Autopilot and Horizon Context to make this easier for Snowflake users.

Cortex Code rebrand

Cortex Code is now CoCo, and she’s coming to your desktop. Snowflake is pushing the envelope by making CoCo available “where you actually work”. This means a Desktop application and Slack integration.

Companies are really trying to move their tools into a central system that has access to everything.

While I think this will be powerful someday, most teams aren’t at a place where they can utilize these features for governance reasons. Focus on the data modeling and semantic layer first, then consider adopting this.

dbt Wizard CLI

This release isn’t surprising given all of the AI hype. dbt Wizard is designed specifically for dbt projects and analytics engineering, but works with other AI providers like Anthropic and Cortex Code. It claims to better understand your dbt projects, especially with model refactoring and tracing lineage.

I already use dbt’s skills with Claude Code, so I’m interested to see if this really performs better than that. Post on this coming soon!

---

Don’t let all of these new tools and features get to you. There’s always gonna be something that’s the “latest and greatest”. Most of the time it doesn’t really change your day to day.

However, I have been thinking about the semantic layer a lot lately. This is one trend that I believe is here to stay with the adoption of AI agents. Data modeling and semantics are going to be the two biggest things to continue focusing on for us analytics engineers.

Right now I’m testing out a new open source tool that leverages your entire data stack to help you build a semantic layer. More on this next week!

What are your takeaways from this year’s Snowflake Summit?

Have a great week!

Madison