All right, good to go. All right, thank you everybody for being here at day two. I'm very excited to have David here from Datadog join us at our new location for TLF in P ark City. We also have Yuka as well. Thank you for being here.
Thanks for having us. Yeah, it's a good location. We're having fun.
Yeah, yeah. The weather's perfect?
The weather's perfect, yeah.
T he vi ew is.
Breakfast was delicio us.
View is outstanding.
Yeah, view is outstanding. Yeah, no complaints.
All right, thank you again. I think a lot of people probably are familiar with Datadog, but I'd love to just hear from you. What is Datadog in the sense of what are you guys achieving from a business outcome solution for organizations? Also, maybe more interestingly, what are the exciting secular drivers that we should associate Datadog to longer term?
Yeah, yeah. Datadog is a modern platform to monitor and observe the cloud workloads, generally that are modern cloud workloads, customer-facing. We have a real-time platform that enables those working in production environments for the most part to see how software is functioning and to investigate problems should they emerge and improve the efficiency. The major driver of Datadog over the long term has been the migration of applications from legacy and on-premise to modern architecture and digital delivery. Companies, whether it be cloud-native startups or the largest companies in the world over different pacings, have been modernizing their infrastructure. Datadog has distinguished itself in having a comprehensive platform that is sophisticated but easy to use and can be used by all the players in doing this monitoring to do their job.
What we've been doing over time is extending the platform from originally infrastructure monitoring to now having a lot of SKUs, including an APM suite, full log suite, digital experience, security, coding, et cetera. The platform has gotten more and more valuable over time. With that, we've grown both the number of customers and the number of SKUs they're using and the revenues.
Yeah, that's excellent. We didn't talk about AI as another secular driver, but I'm sure that'll come up.
We got to save something for later.
Yeah, exactly. 2Q was, I would say, an exceptionally strong quarter for you, acceleration on the top line. I'd love to hear from your perspective what you were pleased with in the quarter and maybe where the upside came from. There are just a lot of things to talk about on the positive side of things. I'd love to hear from you just what stuck out.
Yeah, definitely. At the core of the company, we have been going through an investment cycle both on the product side and the go-to-market side. We expanded very rapidly when COVID happened. We had an adjustment in our end market as customers, the word was optimized. It became more stable, so we had moderated our investment a little bit. We told everybody that we think it's a very long-term and large opportunity. We're going to increase our investment, and we did that successfully. We've been ramping our quota capacity. We've been doing it in geographies across the ecosystem and also been doing it in R&D. In the quarter, we continue to have some marquee lands and expand. If you look at the earnings release, there are a number of very interesting and large use cases where we cross-sold products. We have very significant use of our SKUs.
We save money for clients and added value, and we've been doing that across geographies. On top of that, we've been successfully attaching ourselves to AI. There's an investment cycle going on, which you all know about. Most of the companies that have been successful have been pretty public about it. We've had a growing cohort of software companies that specialize in AI tools, where Datadog has been the preferred monitoring solutions. In the quarter, we did growth as well as breadth. We had growth that complemented our top line by about 10%.
We also had 12 customers get over $1 million, 80 get over $100,000. We're attaching ourselves to that use case. I think another one is we've, over time, given some examples of when we get to certain milestones of products. We've been working on security for some time, and now, this quarter, we've crossed the $100 million mark in security. Those are some of the successes of the quarter.
I think it could get lost like the forest in the trees with investments. I think it's a massive testament to Datadog and your innovation thought leadership in the space that you have this extensive AI cohort of customers, the next generation of technology leaders that are using the Datadog platform.
Yeah, we've always said, and this goes back to that when new technologies form, when we move from development, modern development, when things got containerized, serverless, now AI, that created a more complex set of applications with more of an impetus towards modernization. That's complemented Datadog as long as we're keeping up. I think it is important to note that we are winning in that investment cycle.
Yeah. Those customers have been a beneficiary, and keep me honest, but they've kind of been using Datadog in the traditional sense that other customers would use it.
Yeah.
What's the AI opportunity both to monitor AI applications and for you, you can break it up, but for you to use AI to deliver a better solution?
Definitely. That's a good point because you really have to look at this across the ecosystem. I think the most proximate was that there are a whole set of customers being birthed that are modern software companies that are perfect for Datadog. There are a whole set of enterprises that are beginning to move from training and experimentation into putting large language models in their applications, which Datadog will monitor. We're still early there. The evidence there is the use of our integrations increasing and the use of our LLM and GPU monitoring. Most of the monetization to date has been through enterprises calling out through APIs to these other companies. The monetization has been very much in these tool companies. We're confident that as they mature more and put more applications into production, it will spread. That's the second opportunity. The third opportunity is the Datadog platform itself.
This was a big feature of our DASH User Conference where we announced developments in Bits AI in a number of different ways. The first one was in the platform in service management for production engineers, being able to use large language models to understand the root cause of what's happening faster and then to remediate. That's in private right now. We have a number of customers using it. That was a very exciting part of DASH. It was one of the places where you could hear it in the room. Why? Because service management or the ability to diagnose problems and remediate them is one of the best use cases for AI. We also announced AI Bits in both developer tools and in security. I think that we have the opportunity to have it in the platform and make the platform more valuable to clients.
The last would be, this is the fourth, how do we use it internally? Right now, we're trying to remove the barriers and let our coders, for instance, use such tools as Cursor. We're early stages on trying to see if we can increase productivity and output. That is what we expect in many companies, but we're eating our own dog food, you say, and using it internally and trying to see if we can use AI internally to increase the velocity of innovation.
Yeah, I thought that was an interesting comment. You said the AI workflows moving into inference is mostly happening in the enterprise customers. Are they doing tracing to understand the latency of the OpenAI calls that they're making?
That's early stages. For the most part, enterprises have been, there's been greater investment in training and experimentation. As they're moving into production, that's when LLM monitoring, model monitoring, you know, and integrations will be used. We're seeing usage increase over time, which is a good early sign that it's moving into enterprises in their production applications.
Understood. OK. Bits AI, I agree that resonated deeply at the conference and customers we talked to. I'd say Bits AI 1.0 maybe didn't get as much adoption, but it was tangible in the room, the 2.0 version.
Right.
I don't think the users of the Datadog platform necessarily want to spend, it's not differentiated for them or interesting work for them to detect and remediate kind of just issues. It really resonated when we were talking to customers at DASH.
Yeah.
Given that potential value that you're providing, how do we monetize that?
That's a good question. We are working on right now the economic model where we're thinking about how do we link this to what's happening, like per investigation or per activity. This is true about all of our SKUs. We have floated out pricing, seeing whether it is the right pricing for that type of activity. That's what we're doing now. We don't know the answer. We'll tell you as we get farther along. What we're thinking about is, are we going to have, say, SKUs where you get this capability in the platform and you pay pro-level, championship level, whatever you want to call it? It's not something that we put out publicly, but we do have customers that are paying, that are using, and we're getting good feedback.
Yeah. Coming back to the AI native cohort of customers, and you mentioned the optimization we were all too familiar with back in the day. What are your learnings from that prior cohort of optimizations where they maybe were surprised? I'm not sure, but obviously they had to correct and get profitable. How are you thinking about the learnings from that to make sure as these companies are scaling, that you're proactively setting yourself up to make sure it's not as much of an optimization to go ahead when?
It's a great question. I think if you look through some of the other companies who have faced this, what we've been trying to do is look at the workloads and try to help companies through our account management and our engineers use Datadog in the right way. That would be if they are putting too many logs in or logs that aren't related to real-time production, we'll try to help them send the right amount of logs.
Yeah.
That's one way. Two, we've been innovating the product stack, whether it be Husky or Flex Logs or Frozen Logs, to try to not only suggest changes, but give them solutions that are correlated with their use cases. If you need to have logs frozen or stored somewhere, but you don't need to access them in real time, have a SKU that lets them do that and not pay the real-time price. We've been doing that in the platform. We've also, I think, gotten better about sort of value selling, trying to work on consolidation.
Those involve things like migration credits, longer-term deals, and figuring out how to get a client to find the right value with us, whether it be discounting on volume and increased commitment, having migration credits, having technical account managers attached to them for usage, et cetera. I think those are some of the ways we've been working with our clients to try to evidence value in longer-term client relationships.
Yeah, are you, so Flex Logs seems to be a meaningful way to get customers to control their spend on logs.
Add use cases. It could be, for instance, you need, I think it's more the assignment or the matching. You need real-time, you need access. For this amount of logs, steer it this way and the price, but find other use cases, like use cases to store your logs for compliance. We're trying to match up the price and the technology for the different use cases and, in fact, expand use cases.
Yeah. Is this ending up being a net headwind or a net positive?
It's been a net positive. We're finding that those clients are essentially curating and dividing up, and we're getting our hands on other use cases that we were not maybe able to get our hands on earlier.
Yeah, interesting. Let's talk about security a little bit. We talked about this at dinner last night. My observation on the call, it seemed like there was a tone change from you and Ali on security. I said this last night as well. It's like, from a product perspective, I think you're punching above your weight of where your recent milestone was. Great milestone, $100 million. Your product capability, I think, suggests you could do a lot more. To get to the question, it sounds like you're willing to invest much more in go-to-market on security at this point. High level, what's the rationale for why now invest in the go-to-market and what's the opportunity you see?
Yeah, I think the why now is the product in certain areas has matured enough to be able to win use cases. It doesn't really do any good until the product's at a certain level to expand your go-to-market and have those salespeople or channel partners or whatever get there and not have the product to succeed. We think in Cloud SIEM , for instance, that we're there. The environment with the Splunk acquisition and the product capabilities and Flex Logs and Frozen Logs and all of that is enabling us to have a really credible offering. What we know from how trying to sell security is that it's not good enough to have the best product or a competitive product. You have to reach the buyer, who's maybe a different buyer. You have to go more through the channels because that's where security is.
You have to develop a brand, and you have to help them implement and migrate. Those are all the types of investments that I think you heard on the call that we're thinking about making. It's not going to be a, you know, all in day one, we're going to have another Datadog for security. We're going to try to layer these on in a programmatic way and monitor the success as we go along.
Yeah, makes sense. M&A kind of relative to security, but you can make it broader too. You've had a nice steady cadence of the small talk in M&A. How do you think about that strategy going forward as it relates to security?
Yeah, definitely. I think security is one example. It's really our product roadmap. We just did, for instance, an acquisition in Product Analytics in Eppo, and we did Metaplane in data monitoring. I think we're basically looking at how we can enhance the velocity of the product introductions through mainly technology-based acquisitions. We've gotten good at that. That could be in security. I think we are not adverse to doing a larger acquisition. We're not really making acquisitions just to consolidate customer bases. We'd rather win the customer bases within our platform. If it's the right price, the right team, the right acceleration, we haven't ruled out. We've tended to stick to more of these acquihires or smaller acquisitions.
Yeah.
All about our product. It's all off the product roadmap.
Right.
Yeah.
It is a well-oiled machine, I think, in that process at this point.
We've gotten good at it. Yes, yes.
I want to talk about go-to-market a little bit since that is an area that you're talking about investing. Maybe just on the enterprise side to start, growth has been stable there from a consumption standpoint. You started down market, and you've come up, and you've had a lot of success.
Yeah.
The recent transaction this quarter you called out, that was $60 million TCV, the enterprise space. My observation at DASH is a lot more big logos are consolidated. How do you think about Datadog's penetration from a logo perspective in the enterprise and for the existing customer base, your penetration within that?
We're still, you know, if you look at our penetration, it's still quite low. When you look at the number of enterprise, you know, our penetration, you know, might be somewhere in the double digits. The increased penetration will come from two things. One, we're pretty early in the cloud migration. We still have tons of enterprises that are pretty immature, and the vast majority of their infrastructure is legacy infrastructure. That's a wave we're riding, and we think it's a very, very long wave. Second, what we're doing is we're trying to consolidate. We think we have a very, very compelling value proposition to consolidate the different parts of the observability stack onto Datadog. Don't forget, we didn't have APM or logs when some of these other vendors established their position. By definition, we couldn't have had those businesses. We're going through a consolidation cycle.
I think the third aspect of it is to expand our enterprise motion in a number of ways. First of all, slicing and dicing into key accounts, large customers that will be many years of working on that customer. Then majors, which would be largest customers, and we're working on cross-product adoption and cross-business adoption. Then a hunting group. We've gotten, I think, smarter about that. There's a geographical expansion that we're pursuing. There have been some markets where we arrived relatively later to the game, and we did things in a more centralized way, and we're establishing presence in some of those markets.
Yeah. Should I think of it really closely associated with the go-to-market inve stment priority this year to being enterprise?
Yes, I would say it's prioritized towards enterprise, including channels and things like that. I would say the non-enterprise or SMB is focused on new markets where we didn't have a presence. I think Brazil, India, non-Japan Asia. Our EMEA and our Americas presence is more mature in commercial or SMB.
Yeah, a couple of your competitors support self-managed on-premise deployments. Is that a barrier or something that you would potentially address longer term to really drive full consolidation for some enterprises?
Yeah, I think you're going to have, you're always going to have in the very largest enterprise matching up of the choices for the tools to the business activities. I don't think you're going to have the monitoring of on-premise workloads go away. I think we are thinking about either how to slice or dice the packaging, pricing of the monitoring of on-premise workloads, which may not be as storage or computational intense, and maybe should be priced differently. The possibility of having more dedicated instances, we're looking at it. I don't think it's where the center of the company is going, but we're looking at it as a possibility.
Right. We have a couple of minutes left. I want to survey the room, see if there's any questions out there. All right. Here's one.
How do you look at the competitive environment on open source?
It's been the same as it's been for a while. There's been an impetus towards buying a system like Datadog, meaning the revenues in open source have not been growing as fast as Datadog. There are always some places that want to combine open source with the system or want to try to insource. I would say it hasn't become more intense in the last few years.
Yeah. OK. Maybe last one, David, financial question. Just margins are compressing a little bit because you said you wanted to invest in capturing the opportunity.
Yeah.
How do you think about the longer-term growth versus margin trade-off?
Yeah.
Further out?
Yeah. I mean, we stick to what we have said in our investor day, which is our long-term target is 25%+, and with free cash flow 200 basis points to 300 basis points higher. We've already proven we can get there. I think this is really a situation where some companies are still proving they're economical. We've proven we can become very profitable. For us, it's identifying good investments that we think can compound the top line. There are a lot of them out there. The way we're looking at it is being disciplined, prioritizing them with trying to meet our obligations in profitability and continuing to improve. We have a profitable business, not leaving on the cutting room floor growth opportunities. That's how we do it. It's a balance. I think we've been good at it, and we'll continue to balance those two things.
Two quick ones.
Yeah.
As a CFO of a software company whose job it is to create software, how much software in three to five years, pick your time frame, is written by AI? Question one. Question two, if you see productivity gains from Cursor tools and stuff like that, do you ring the register on those savings, or do you just deliver more product?
I think we've been clear we're going to deliver more product right now. I think that might have to do with the fact that the productivity gains are there, but not proven out completely. I wouldn't be a CFO unless I wanted to have metrics on productivity, and I wanted to match that up with the demand for product and product enhancements. I think we'll eventually get to some potential savings, but I think in the near term, we'll try to use that to sort of get our products out the door faster. I'd like to see, you know, once everyone gets comfortable, I'd like to see metrics and proof. I'm my kind of person and see where we go from there. I think there's opportunity, but the opportunity still needs to be realized.
How much code writt en by AI, best guess?
I don't know. That's a good question. I don't know.
All right, we'll come back next time and look better at it.
I do not know.
All right. Let's give a round of applause for David. Thank you so much.
Thanks a lot. Thank you.
David, appreciate it so much.
Thanks a lot. Thanks for inviting us.