What is the business model of dbt Labs as an open source company and how do they differentiate between their open-source offerings and paid services?

So dbt Core is a language or framework. It's SQL-based with some extra bells and whistles, like configuration. We have Jinja and macros so that you can be more productive with SQL than you could otherwise. 

We’ve built a compiler that translates your dbt SQL code into instructions for your data warehouse to create data assets. That's dbt Core. It's open source and will always be so, because it's a language, and it's really valuable for companies to have a place to describe something as critical as their business logic in a framework that's not proprietary. dbt is a standard, and it will always be freely available for our users. 

On the cloud side, we build a lot of proprietary software that complements the dbt workflow that helps customers accelerate their time to getting up and running and improves their data posture and velocity.

The cloud product has an interactive development environment, so you can create your dbt models more easily. It has CI checks, so, when you're trying to promote your code from development into production, you get a staging environment that automatically tests your dbt code and data quality. 

We have a scheduler that powers building models so that your tables can remain up to date and fresh, and you don't have to maintain that infrastructure yourself. 

We’re launching a semantic layer which will power your business queries, and allow you to have good governance so that you can write your business logic once and have consistency wherever it's queried or consumed. 

We also host your documentation for you. The cloud product, what we sell, supports the full workflow around the analytics engineering practice, supercharging the open source language and framework.