Databricks Doubles Down on AI Efforts to Take on Rivals

You and I have talked a lot about what's happening in generative AI and the people that are building the lands. And at a time you were agnostic, you said, okay, there's loads of great lambs out there. We can pick and choose. You've changed strategy. Why? What? We're still agnostic. You can still use all the different large language models on Databricks, but we built DB X and we open sourced it. Why? Because our strategy is that most of these enterprises have proprietary data and they want to hav

e their own large language models that understand that proprietary data that we call that data intelligence, understanding their data. Other companies are building general intelligence. You can ask it about World War Two or whatever you want, but it doesn't understand the data that these enterprises have. So for that, we have an open source model that you can customize and can own the IP so these enterprises can own it. So companies like Rivian or BLOCK can build their own large language models,

which they are with us that then understands their data and helps them become competitive. So let's simplify this. Let's run with the Rivian example of this companies. You know that I know very well. You are making a generative A.I. tool for them where within the confines of Rivian, it can give answers based on a prompt specific to rivian. Yeah, the company and its data just explain the basics. Yeah. If you take Rivian for instance, there are models that are custom on all the data that they hav

e can switch lanes, can, you know, look at the cars in front of it. They can even optimize the energy consumption of that car, right? Custom model on their data. It's very important for them to be competitive against their competitors. Another company block formerly in square large language model. And then you can interact when you register on their devices and you know, you're interacting with a square app, It now can understand you can just speak English or any other natural language based on

these large language models. So this is key to these companies. It's not general intelligence, it's data intelligence on their data. Ali, There are loads of founders and CEOs of publicly traded companies that come on this program. They know exactly what Databricks is. You would say 70% of Fortune 500 companies use databricks. I get it. There is a ginormous body of people in the world that haven't got a clue what Databricks is and why it's important. Explain this new land in the context of what D

atabricks does. As simply as you can. Yeah. So look. What has happened over the last ten years is that every company on the planet is moving and outsourcing the IT infrastructure to the hyperscalers, to the cloud vendors. It'll be us, Google, Microsoft. They also all need a data platform. That means data that they bring in so that they can create, you know, Tableau charts and Excel charts, but also do predictions on it to predict what's my revenue going to be, which of my customers are going to

turn next, when should I replace my equipment? So basically data and AI. So we are the go to company. If you want to democratize data and AI inside of an enterprise, you probably use the hyperscalers for your infrastructure of I.T. But then for data and A.I., we do that for you. We help you do both analytics. That's backwards looking. What was my revenue last week, but also forward looking? What's my revenue going to be next week? So that's what we do. We democratize that to enterprises. Ali, yo

u're an incredibly competitive person. You are doing this to make Databricks more competitive. I'm assuming principally against Snowflake. This is the first time you and I have had a chance to talk since Frank Sleeman stepped back, stepped down and they brought in a more product focused leader is my read. What's your reaction to all of that? What do you make of it? Yeah, I mean, I think we put a lot of pressure on them and they realized that, hey, AI is important. So Snowflake basically was not

doing any AI whatsoever. They're actually a great company with a lot of respect, a lot of respect for Frank, actually, who I think that an excellent job with that company. But they're primarily a data warehousing company. They are offering super important technology, but it's used to ask questions about the past. You know, how did this product do last week? It's not predictive. Yeah, it's not predictive. So and we've been doing, you know, I for the last decade. So I think it shows that, you know

, the puck is going towards A.I.. I mean, we see that which is generative A.I. models with DB X with what everybody, you know, every CEO I talk to now, Fortune 500 company will tell me I super critical for our strategy. We think that actually in our whole industry data now is how we're going to become competitive. Help me do that. So of course, it makes sense that a lot of vendors out there are kind of pivoting now and going towards say.