Just to start with the optimizations point, from my view, Snowflake optimizations are a fairly obvious thing to do. Observability, observations, optimization. Same thing for Mongo. It feels like a much higher degree of difficulty. Maybe you have some dev test environments that are running too long, and you spend those down. But what can you do as a customer to really optimize Mongo spend? Because that seems like a lot of work.

It has not been frequent that something comes along in this space that allows you to organize the environment a little bit and keep it from getting more chaotic. It's been more chaos than less over the past several years. But this is a thing. I would say thinking about multimodal databases has been a slight negative on Mongo, because unless you have JSON-heavy workloads, there may be another database that's good enough for the little bit of JSON that it has in it. I think that's where some of the optimization comes from.

Okay, that gives me a good feel for how you're coming out on them. Let's turn to Snowflake. The investor narrative for Snowflake has been probably the worst I've seen for a very long period of time, but I'm wondering if that's actually carried over or if that's reflective of customer sentiment. And I guess the best way to frame this question is, is Snowflake getting their lunch eaten by BigQuery, Fabric, and Databricks?

AI hasn't been as explosive as we thought, but it's an essential part of most new workloads. We just haven't had many new workloads. But now, I'm starting to see activity pick up, and everything is about what we can do with GenAI. There was a lot of thinking in the first half of the year, and now we're turning that into pilots and occasionally into production. That's going to favor Databricks. It's really down to Snowflake and Databricks, with some BigQuery, some Synapse, and some Redshift. But Snowflake and Databricks are the two battling it out in most enterprises, and Databricks is mostly winning.