Snowflake Summit 2025: Key takeaways for modern BI teams

This week, I traded my tiny Alpine town in Northern Italy for Snowflake Summit in San Francisco. The summit brought thousands of data professionals together for four days of talks, networking, and enough fake snow to make me homesick.

‍

My first insight? Conferences hurt your feet.

‍

But more interestingly, I learned a lot about the future of the data world. Here’s what stood out:

‍

AI is everywhere, but semantic layers are doing the real heavy lifting

‍

You couldn’t attend a talk without hearing “AI”. But AI in BI doesn't work without a solid semantic layer - it’s what gives structure and meaning to raw data and turns AI from a buzzword into a genuinely useful tool. Our semantic layer is built on dbt, meaning you can go from raw data to transformations to AI insights all in one place. Nice, clean, purposeful analytics.

‍

We also teased version 2 of our AI analyst during the summit, and the response was fantastic. Powered by the Lightdash semantic layer, it integrates seamlessly with our self-serve experience. People stopping by the booth were genuinely impressed by how natural it felt.

‍

Data is expensive

‍

Data is powering the AI revolution, which means more storage, more transformation, more tools and more costs. Nearly every other booth promised to cut your Snowflake costs, all while the tools people are actually using keep getting more and more expensive.

‍

This is exactly why we don't price per seat at Lightdash. Attendees we spoke to were pleasantly surprised to learn they could scale their BI deployment and put AI in a user’s hands without watching their budget explode.

‍

dbt runs in Snowflake workspaces now

‍

Snowflake also announced that you can now run dbt directly in Snowflake workspaces! As both a dbt fan and Lightdash advocate, this was great news. More dbt integrations, simpler dev experiences, and broader adoption are all positives in my book.

‍

Embedding is almost as hot as our sauce

‍

There was loads of buzz around embedding dashboards and analytics into products. Our new React SDK is hitting the mark, and it sparked some great conversations with folks building data-rich apps. It’s always exciting to work with companies that want to put data directly in the hands of their customers.

‍

Devs still 💜 Lightdash

‍

Lightdash was built by and for developers. The code-first approach to BI that we were founded upon continues to impress. It really is the perfect tool for Analytics Engineers, and conferences like this help confirm that what we’re building is solving real problems that BI developers experience.

‍

Hot sauce wins the swag wars

‍

Forget branded socks. We bought three new Texas-made hot sauce flavours this year. If you missed out, look for us at our next event. Are you brave enough to sample “Data Inferno”? It lives up to its name.

‍

Final thoughts (and a home run)

‍

As a remote team, it was awesome spending time with some of the other folks at Lightdash: Ian, Katie, Marshall, Jake, and Hamzah. Most days I’m working from a tiny town, so the energy of SF (for a few days) was a fun change of pace. The Giants won the ball game at the bottom of the 10th inning which helped the vibes!

‍

‍

The next stop for me is Big Data London in September, where I'll be talking about Analytics Intelligence Engineering and how Analytics Engineers are perfectly positioned to build context for the AI agents of the future.

‍

Want to see what all the fuss is about? Book a call to try Lightdash for free and experience the future of code-first BI yourself.