Business Breakdowns
· Alan Tu
[00:00:00] S1: Well, maybe just to start with some quantitative framing. So, Databricks, they're now over $4 billion of ARR. They have disclosed that about a quarter of it, $1 billion, is AI-related revenue. And so, AI has already become a very large part of the business. But I think beneath that, there are different ways to perhaps slice and dice the impact of AI on Databricks. And I think for me, one of the things that I've really liked about Databricks as an investment is that there are multiple ways to win. Starting with the core data processing piece, there's almost a consensus understanding within enterprises now that you don't have an AI strategy without a data strategy. Everyone recognizes, of course, the model providers are doing what they're doing, and every generation of models is getting smarter and smarter. But at the end of the day, if you don't have good, clean, well-cataloged data, the models can only do so much. One of the ways that AI has really benefited Databricks's business is that it's just created a tremendous amount of prioritization and awareness of the importance of the core product that Databricks has always provided. In my mind, that is a durable tailwind. That companies will need to do, have always needed to do anyway, that is actually not dependent on whether we achieve AGI or not, or what does the next OpenAI model do. The reality is that as long as there is a general belief and understanding that AI is important, there will be a driver towards more data engineering and data processing. That's a general tailwind for Databricks that I think, when you think about Databricks the business as an investor, it actually paints a picture of a more durable growth trajectory that is perhaps not as spiky on the upside, but also not as volatile on the downside either in the case that sentiment around AI changes.
· Matt Reustle
[00:00:00] S0: That makes a lot of sense just in terms of a heuristic at a high level for thinking about AI within the business.
· Alan Tu
[00:00:00] S1: Yeah. And there's probably a couple other ways to think about AI's impact on Databricks. Another way is actually the fact that there is this huge cohort of AI-native companies, including the largest AI labs, that can and do use Databricks as well for themselves internally. This is something that companies in the public market also talk about, is that there's, "How is AI actually impacting product and use cases?" And there's also the fact that, "Are you part of the stack that AI-native companies are utilizing?" And Databricks is very much there. And then I think the final part of AI's implication on Databricks as business is actually product. One of the big picture bets that Databricks is making, going back to this idea that they have a DNA of having an opinion about where the world is going and where can Databricks add the most value, it's really around the idea that in the future, AI and LLMs have already proven, even if the models don't get any better than where they are today, the ability to automate more work. And when you think about how big of a TAM that is, it's probably just as infinite as the TAM that we talked about initially around data. So, what Databricks is doing is building products that, in the same way that they came out with MLflow that helped that whole process of a data scientist building a machine learning model, they're creating an entire stack with products called Agent Bricks and LakeBase that all together will help enterprises build their own agentic applications to automate specific use cases and actually automate labor and work, which is just tremendous amounts of ROI.
· Matt Reustle
[00:00:00] S0: Yeah. And I'll reframe it maybe, and you can tell me how accurate this is versus inaccurate. Most of what we talk about with agentic is oftentimes just shrinking the context, giving it very deep context in specific tasks, and therefore the quality of the response is going to be much stronger. It's not going to be pulling random places on the web. Databricks obviously has the most rich data to use in terms of informing those models and therefore can partner with businesses who might want to develop those agents.
· Alan Tu
[00:00:00] S1: Yeah. So, to your point, part of what the industry is realizing is that there are techniques that are important to be leveraged to build effective agentic applications. For example, RAG, which is retrieval-augmented generation. There's a whole set of processes around enabling that, with vector databases, for example, embeddings. So, Databricks' offerings there. There's also a very important part of building agentic applications is around model evaluation. So, because of the very unpredictable nature of the large language models, it's not quite as easy to always know exactly, "Is our application or agent acting the way that we think it should?" And so, there's a whole set of technologies around model evaluation, around being able to actually quantify, " How are these models behaving? Are they doing what we think that they should be doing?" And Databricks again is building products around that. In the same way that it's parallel going back to the machine learning era where you would go through a similar process around evaluating models there. You can get a sense of the fact that while there's a lot of focus on the core large language models, if you actually want to build applications in production, there's so much around and beyond just the model that is really where I think Databricks has a strong right to win.
· Matt Reustle
[00:00:00] S0: Yes, the infamous question of where the value might accrue within the layers of the AI ecosystem is quite interesting. On the opposite side of the equation, there's been this interesting, solid, good relationship with the cloud infrastructure providers. AWS is there. Does that change at all with AI? Because to your earlier point, the data cleanup and all of that becomes even more important. I have personally been able to clean up unstructured data into structured data much more cleanly with AI. Different story when it comes to me versus doing this at cloud scale, but what's your view on where that relationship, where the cloud providers haven't necessarily moved into this category, does that change at all? Is that a risk? Is that something that you think about at all?