Sonak Kolar: This is Sonak Kolar on for Mark Murphy. Drew, over the years, Dropbox has kind of created this immense repository of content, I think, over $1 trillion now. So as we think about this Gen AI opportunity and the need to kind of pull high quantities of content into the LLMs to train them, can you just help unpack how Dropbox’ scale differentiation is resonating in the market? And maybe how many more deals or partnerships you’re kind of getting pulled into specific to AI because of the centralized content repository?
Andrew Houston: Sure. Well, I mean just a couple of clarifications. So I mean we don’t train foundation model. So when you look at like what OpenAI does or what Meta does with Lama basically taking the whole Internet and training foundation models on it. That’s not really — that’s something we can kind of get as a service or in Meta’s case it is even open source. So we’re sort of beneficiaries of that. And then there’s also sensitivities around — we take our responsibilities around privacy super seriously. And so while there’s a lot of benefits to having this kind of scale of data. We’re also mindful that — or you have to be super careful if you’re training any models on people’s private information without complete transparency and control over that.
And we find that like to build a good AI-powered experience, you actually don’t need to train models on people’s private information. But the scale advantage comes from being able to — I mean, there are other aspects of like the engagement with Dropbox or when you use our service factor, we have a big — such a big audience the fact that we have all these technical investments in like understanding content as there’s new capabilities that come online from the research frontier around multimodal models, being able to handle audio and video, better. Those can translate to big advantages for us because we have a huge — we do have such a huge repository of that kind of content. And we’ve been adding ML capabilities both in the pre LLM world with things like being able to automatically scan a document with your phone and sort of auto correct or do image or OCR see text and images and search on that.
There’s a lot of similar things that we’ve been doing, and we’ll continue to do around applying AI to your content, whether that’s transcribing things really easily. We already have that in a number of our products. being able to organize your Dropbox for you, being able to understand the images and video and audio and your Dropbox better. So there are huge advantages that we have across the board, both from the scale of our user base and then a lot of the technical investments we have in content. And then a lot of the kind of technical investments that led us to have such — to optimize our margins and performance on the storage side, a lot of that — those kinds of technical problems translate super well, to large language model inference and bringing a lot of that in-house over time.
Sonak Kolar: Got it. That’s super helpful. And then I just had a quick follow-up related to the video component and the fact that you shared that video is the fastest-growing content type on the platform. Is there any way to perhaps size a long-term opportunity that the surge in kind of video or short form video is presenting both to Dropbox and then the maybe how you’re thinking about the future innovations around replay where you’re kind of embedding some of that AI capabilities beyond Dash?
Andrew Houston: Sure. Well, I mean, first, it starts with our customers. I mean, one of the most popular — popular things people do on Dropbox is create different kinds of media or collaborate on different kinds of media. I mean, video is one example, but also audio photos. And a lot more. And we’re really differentiated there because of the fact that Dropbox has supported big files so much better than competing alternatives. And that we become kind of a de facto standard in the creative community. That’s a really strong foundation where we can — that we can build on. So things like Replay come from observing just folks like actually often like collaborating on Dropbox and working on a video, but then also having to have like a text threat open saying like, “Hey, it’s at 4:15 that little — can you get rid of that little yellow balloon in the corner.
Like there’s all these collaborative workflows that were like, hey, this should all just be integrated into the Dropbox experience. And as things like AI have been coming online and the example I mentioned earlier is like, hey, you should just be able to like review a video and the text of what the person is saying should just be there. And you should be able to either view or collaborate or edit the video in a more natural way. And these multimode models are going to open up all kinds of functionality like that. So a lot of it just starts with like — all right, what are our creative audience is already doing on Dropbox. It’s not hard to see all these ways we can make that experience better. And then AI opens up all kinds of technical new doors, to connect the dots between what our customers are doing and making a much more streamlined experience.
Operator: [Operator Instructions]. Our next question will be coming from Brent Thill of Jefferies.
Luv Sodha: This is Luv Sodha on for Brent Thill. Maybe first to start with you, Drew, just wanted to ask, maybe at a higher level, if you could talk about the core FSS business, one of the investor concerns which it sounds like from a bundling experience, there is a fear of commoditization in your end markets. And so how do you think of that core FSS business growth, say, 2, 3 years from now? Do you think that, that business is still growing? Or is it declining? Just any color on that would be super helpful.
Andrew Houston: Sure. So I mean we still see a lot of room to keep optimizing both in our teams business, and I gave some examples of that earlier. We do — we hear things about commoditization and we certainly are operating in a competitive environment. But at the same time, we’ve been able to keep growing our business at $1 billion — or more since we went public and ARPU has been going up. So there’s a lot of health in the fundamentals of the business, too. We see like files as an evergreen need, right? Like Dropbox is mission critical for our customers. And then I just talked about the creative community and folks who work on large files or content, Dropbox is specially mission-critical. And when I talk to our customers, they remind me that there’s no shortage of new things we could be doing or doing better.
