Snowflake Summit, Unfiltered: One Week, 20,000 People, And A Whole Lot Of AI

Snowflake Summit 2025 is massive. Twenty thousand people. Hundreds of sessions. A flood of product launches, partner news, and AI talk at every turn.

To make sense of it all, we asked Stipo Josipovic—our Senior Director of Product Management and a seven-year Sigma veteran—to be our eyes and ears on the ground. He’s been to six Summits. He knows what matters, what doesn’t, and how to spot the signals in the noise.

What follows is his personal journal: an unpolished account of the moments that stood out. What he heard in the halls. What he saw on the floor. What people were really talking about once they stepped away from the stage.

This is the inside view.

Monday, June 2

9:42 AM
Every booth says AI.
I did a full loop right after doors opened. The floor is massive, loud, packed—and covered in AI signage. A few standouts: Metadata’s going big. Catalogs aren’t dead, just evolving.

3:23 PM
The NYSE showed what good looks like.
In a packed session, their team shared how they use Sigma to drive 70% org-wide adoption—managing more than 7 petabytes of data and running over 150 million queries per month through Snowflake. Afterward, an attendee stopped me in the hallway. Thirty minutes later, they were asking our team about pricing—no POC, no push. When tech meets trust, things move fast. 

4:07 PM
People aren’t asking “Why Sigma?” They’re asking what we can replace.
Met up with Kyle Lange and Athan Hsiao from sales, and Greg Bonnette, one of our field CTOs, at the booth. The shift is clear: folks already know Sigma. The questions are pointed—Power BI, Tableau, even Looker. It’s not just interest anymore. It’s intent. One person even said, “I’ve been trying to do this in Power BI. It’s not working.”

5:00 PM
Data apps get real reactions. People get it.
I checked in with Ben O’Neill on our sales engineering team after a few hours at the booth. Almost everyone he talked to had heard of Sigma. Most weren’t looking for an AI pitch—they had other problems. But when he showed them data apps, their faces lit up. “Didn’t matter the use case. They saw the power.”

6:57 PM
Fred Studer, our new CMO, is on week eight and already losing his voice.
He’s worked 500+ events. Said he’s never seen this kind of traffic. “Five people deep for demos,” he told me. The thing people kept calling out? Not just the AI—but how simple it is to use. 

‍

Tuesday, June 3

11:08 AM
Snowflake’s vision is expansive—and accelerating.
Just out of the keynote. In 90 minutes, they rolled out a staggering range of launches: catalog enhancements, ingestion tools, AI agents, warehouse improvements, and new ways to write and automate SQL. It was all framed around one core theme: I want. As in, I—the customer—want a future-proof data foundation. The message was clear: Snowflake isn’t just shipping fast; they’re structuring it around what customers actually need. The orchestral cues and live demos helped make it memorable. Now the big question is: how quickly can it all arrive?

12:16 PM
Partner of the Year. Third year in a row.
The booth was buzzing. High fives, hugs, a real sense of momentum. Recognition like this doesn’t happen by accident.

12:21
Big launch today: Semantic Views and Cortex AISQL.
Semantic Views lets teams define shared logic in one place—no more copy/paste metrics or one-off definitions. Cortex AISQL takes that context and turns natural language into real, working SQL. It’s a shift. Not just in what we can do, but in how people are talking about Sigma. This isn’t “someday” AI. It’s shipping—and people are showing up to see it.

4:17 PM
The market is validating what we’ve believed for years.
I debriefed with Kyle Herold, our principal for data applications. He’s four months into Sigma, and already he sees it: customers want to break out of static tools and spreadsheets. They’re trying to fully activate Snowflake, and they need a front end that meets people where they work. One company even told a room, “We’re like Sigma” That’s a shift. We’re not just a product anymore—we’re a reference point.

‍

Wednesday, June 4

11:45 AM
AI is everywhere—but clarity is rare.
Every analytics vendor is pitching AI. What’s missing is trust. Buyers don’t just want answers—they want to see how those answers were made. That’s why the “show your work” approach is resonating. We built that into Ask Sigma from day one. Metadata is also having a moment. Between Horizon, Semantic Views, and Iceberg investments, it’s clear: the future of AI starts with accessible, structured context.

12:48 PM
Data movement is having a main character moment.
Snowflake’s Openflow announcement raised eyebrows—but not alarms. What stood out in conversations today is how critical data movement has become in the age of AI. It’s not just ETL or reverse ETL anymore. It’s bidirectional access to every system, every file, every source. That’s the foundation AI needs. The shift is clear: this isn’t about replacing data movers. It’s about recognizing them as essential infrastructure for what’s coming next.

5:40 PM
Snowflake isn’t just shipping features—it’s assembling infrastructure.
Closed the day with the Infostrux team, who built the LLM-powered assistant demoed in the Whoop keynote. It’s not just chat in Snowflake—it’s orchestration between agents. Governed data meets best-in-class models, routed through a unified interface. “It’s not ChatGPT versus Snowflake [on a worse model]—it’s ChatGPT with context.” The shift is subtle, but foundational. The future isn’t a better prompt. It’s a smarter system.

‍

Thursday, June 5

10:18 AM
Data products are the new connective tissue.
The floor started slow, but conversations picked up fast. I caught up with Shawn Namdar, our Sr. Technical Alliance Manager for Snowflake, to talk shop. One thing’s clear: the marketplace continues to be important—not just for datasets, but for connected apps and operational tools. “Data product” is becoming the umbrella term for anything built to work inside that ecosystem. At the same time, the metadata wars are heating up. Every catalog vendor is leaning in, making it obvious: AI needs context, and the fight is on to supply it.

12:08 PM
There’s a lot to take in—but some things stand out.
‍I ended the day at Dev Day, watching early-stage teams pitch apps built directly on Snowflake. AI tools, security workflows, vertical use cases. It’s early, but it’s a clear signal: Snowflake isn’t just expanding its platform. It’s creating space for others to build.

Across the week, a few patterns held up—AI is the headline, but trust is the subtext. Metadata is rising in importance. So is data movement. The landscape is shifting—and I’m paying close attention.

‍