GitLab Inc. (GTLB) Earnings Call Transcript & Summary

Afternoon session, Day 4 of Morgan Stanley TMT Conference. I'm Sanjit Singh. I run infrastructure software at Morgan Stanley Software Research. I'm super happy to have the management team of GitLab. We have CEO and Founder, Sid Sijbrandij. Thank you Sid for joining us at the conference, I think for the first time. And we have Chief Financial Officer, Brian Robbins. Brian, thank you for joining us at the conference as well. Let me get through some disclosures, and we'll get right to it. For important disclosures, please see the Morgan Stanley research disclosure website at www.morganstanley.com/researchdisclosures. If you have any questions, please reach out to your Morgan Stanley sales representative.

So just to start the conversation, I've been following GitLab for a long time. And I think what's always impressed me is about the boldness of the vision, right? It's always been thinking about the software release cycle holistically. Was taking a platform approach to developer tools and now increasingly dev sec ops, was that always the vision? Or like when you started, were you looking for just solving a singular pain point? And how did that go from that to this very bold and ambitious vision to tackle these big markets?

Yes. We were looking to solve a single pain point version control. And then, Dmitriy, my co-founder, never wanted to upgrade another Jenkins installation again, and he made another product called CI without asking or a permission. And then at some point, 2015, [ Kamil ], an engineer in Poland said, "Hey, I think we should combine that 2. And Dmitriy said, obviously, that's a bad idea. We need sharp tools that our people can combine. I said that's a really bad idea. Our customers want to combine tools. They don't want everything integrated. And Kamil persisted, and he said, "Well, at least it will be way more efficient for us to just maintain one code base. And we said, "Okay, well, we'll do it. But the results really surprised us. Not only was it better for us, but it was a lot better for our customers. Today, when customers switch to GitLab, they get an average speed up of 7x -- 7x faster cycle time because they don't have to hop between 10 different applications. And that's been the power. And since then, we had unlocked kind of that knowledge, and it was obvious. We shouldn't combine 2 things. We should replace like the 10, 15 point solutions that company use. Sometimes a big bank like 170 different point solutions that we can consolidate on a single platform.

That makes a ton of sense. And it seems like software today is going to more platform projects, particularly coming out of the last cycle into this higher rate environment as this new normal because enterprises are focused on efficiency. Thinking about the last year, I mean, growth has been impacted like many software companies by the tighter budget environment. But when I look at your Q4 -- your most recent results, there's a lot of things to like, right, like revenue growth of 32%, net retention rate actually improved to 130% from 128%, 52% growth in $1 million customers, 37 for the growth at 100,000 customers. And I think Brian even commented that there was some sort of normalization in the buying cycle. And so the question is, like in the context of all of those, call it, green shoots, if you will -- on the call on Monday, you noted that guidance would be "less conservative" going forward. Can you share your thinking of what we can expect as upside as far as to guide? And sort of why did we -- why did you have to frame it in like the context of more concern versus the past couple of years?

Thanks for the question. This has been a question that has come up a lot since earnings. I appreciate you asking it. One of the things that GitLab is we're a transparent company, and we've always been transparent. When we actually -- for the first 2 years as a public company, our average beat has been about 7%. And so as we're going into our third year, we really wanted to be more prudent in our guidance, and we want to let people that we'll get closer to the pin, and we'll be less conservative than we've been historically on our beats. We do have a number of growth vectors that we talked about in the company that are long term in nature that we're super excited about. We have the premium price increase. We have the agile planning SKU. We have GitLab Duo Pro. And so there's a lot of growth features in the company that we're excited about.

That's an important framing of the guidance. Some of the other themes, right, and then we're going to have a conversation, a pre-in-depth conversation about how AI is impacting the software release life cycle. But let's talk about Duo. That's something that you highlighted on the call. Can you explain to the audience what Duo is and how is it being used to accelerate and automate each workflow that makes up the broader software development cycle?

Yes. So Duo is our suite of AI features, and we address more work streams than anybody else. The recent our report show that we're by far in the lead in addressing use cases. Typically, when people say AI and DevOps platforms, they think of cogeneration, but there's a lot more. We've got 15 other things that we address with GitLab. You can imagine a developer writing code, it's typically 25% of their time. We help them with the other 75% of their time to summarizing issues, doing all kinds of other things. But we also -- for every developer, that's typically an operations or a security person, and we make their lives better as well. So we explain vulnerabilities with AI, we suggest fixes, we help with planning, we help with deployment. So AI is not like a single feature. It's all these features together to make people more productive. And because we're the broadest platform, we are planning and creating and securing an operating software in one platform, we can offer more AI features than anybody else.

Awesome. And so Brian, to you, like with Duo, I think you have Code Suggestions on the way, you have pricing up for these solutions. Where are we in terms of the customer spending intentions? You sort of addressed a little bit in the previous question around guidance. But with respect to sales cycles, customers downsizing, churn, to what extent is that getting better going into 2024?

Yes. We talked a lot about customer buying behavior patterns have changed. And so sort of pre-pandemic about 1.5 years ago, 2 years ago, when a procurement department would come in and buy, they would buy as much as they could for projects that were going to be funded, headcount that was going to be funded, they would try to negotiate as big of a package as they could and get the best discount with the ramp deal. Through the pandemic, when there is a number of different layoffs, the buying behavior change where they would come back to you on a much more frequent basis, and they would buy just for what's needed, and they would try to come to you multiple times a year. And so we commented in this last call about how the buying behavior has more normalized. We saw churn in contraction levels improved to what they were like 6 quarters ago. And so we're super happy, especially with the large enterprises on the performance that we had this quarter. RPO grew 55% year-over-year, which is really a testament to the large enterprises and some of the multiyear deals that we signed in the quarter.

And as we think about on the -- we're talking about AI and hopefully going into a new innovation cycle, you guys have a seat-based model. I like -- sort of connect the data points, the hyperscalers are doing a little bit better, they're seeing cloud optimization stabilize, is it just sort of a matter of time because if we want to release AI applications we have to release more software. Is it just a function of sequence that ultimately, as the hyperscalers start to see new workloads coming on, that drives more software development projects and you ultimately sort to see that improved demand in results.

I personally think that as AI makes creating software more affordable, that we're going to see a lot of demand. Like if you make it easier to make software, there's going to be more people who make software, not a few of them in my opinion. And we're super happy with our partnerships. With the hyperscalers we mentioned in earnings that revenue through the hyperscale has increased 100% year-over-year.

Going back to the first question around taking a platform approach. When I look at this market, it seems like the software delivery pipeline, there's a couple of things going on. It seems like software delivery pipeline is in process of moving to the cloud, number one. Two, we got AI being infused across the various workflows. And three, customers are more focused on productivity and efficient than they ever -- they ever have. And so the question is the time for a platform now and are we expecting to see an inflection in customers take a platform approach because to me, this market kind of felt like security. It's been -- you overlay, you have the software release cycle and there's 20, 30, 50 vendors solving for each one. And so is now the time where we should expect customers to rethink how they're delivering and securing software.

If you look at your 3 factors, I think it adds up to move into a platform. The hyperscalers partner with us because they know if they first standardize and GitLab at a customer, they can move to the cloud faster. If you look at AI, the broader the platform, the more AI features you can bring to bear to help people. And productivity efficiency-wise, customers get a -- kind of a 4-layer return on GitLab. They see some software licenses, we don't have the integration costs of bringing all the different point solutions to get it. Their people get more effective. But most importantly is that increased cycle time, like faster cycle time means more progress and your most important projects. Our most expensive offering, GitLab ultimate pace for itself within 6 months.

Makes tons of sense. And speaking again about other things going on in the market and potential catalyst, you have one big player in Atlassian who has this end of life of their server deployments. When you talk to your customers and if you can take us through behind the curtains and the eyes of a decision-maker who owns the software release cycle, when they have to think about, okay, I have to do something with my [ GitLab ] you either have to take it to the cloud like a [indiscernible] or maybe seek all the alternative solutions. Is that creating a funnel, if you would, well, or a decision point where people have to think about a broader architecture decision.

Yes, being forced to make a decision. People took stock, like what did we want? Like how is the spend going to evolve on that? Do we get the same functionality, but also how should our process look like? And with GitLab, a lot of it checks that were manual before are now automatic. So it's less manual work, and that attracts people. We had a leading 3D design firm, not only buy a ton of GitLab ultimate licenses, but also by enterprise agile planning licenses for all their salespeople. The people not directly involved with DevSecOps, they're part of the software planning process. So the biggest benefit when they move to GitLab is that the planning happens on the same platform as the creation. So it's more up to date, there's less manual work and you can move faster.

As a follow-up to that, just came from -- you answer the question. From a sales motion process, are you targeting some of those Atlassian customers who are looking at that deadline that was in February and sort of prospecting them? Like what's been sort of the -- is there a specific sales playbook to address that opportunity?

There is -- to address that immediate opportunity at the kind of depreciation of their server product. And I think more importantly, I think it's going to be something that keeps on giving because the opportunity to consolidate point tools into your platform, to bring your planning together with your software creation, I think that's a good idea today, but that will be a good idea 5 years from now. So that's why we're really excited about enterprise agile planning for the long term.

Awesome. Brian, it's been a focus for investors around the price increase that the company launched on its premium offering last year after, I think it was 4 or 5 years and you guys have at least 400 new features and capabilities. So it was probably the right time. That pricing went from $20 per user to $30 per user, right? And so in terms of -- on its face, it looks like a big 50% price increase. what's been like the effective yield on that? I mean, in terms of what's been the reaction to the buying behavior, right? Because there's a list price increase and then customers can decide, hey, maybe we have fewer users on premium, maybe we upgrade to Ultimate. What's been the like behavioral response to the pricing change?

Yes, absolutely. When we announced a price increase, there was a lot of different calculations for how much price increase is going to be. So we've done a lot of time educating investors and analysts on what the impact will be. When we model it internally for FY 2024, we actually did better than our expectations in our internal models. And so we're happy with the results. We talked about the impact for the price increase in FY '20 and '25 is going to be $10 million to $20 million of incremental revenue. I think if you look at the price increase, it's been 5 years, over 400 features. We've invested hundreds of millions of dollars on the platform. And so the price increase was just justify the value that we're delivering to our customers. We have cohorts back to 2016 that are still expanding with us today. There's not many other businesses with the same product, same price point for over 8, 9 years, they're still expanding at the same rate that they're expanding from 2 years ago.

You've seen a lot of adoption of the ultimate tier. To what extent has the price increase on premium sort of created a nudge in incentive to upgrade to Ultimate?

The Ultimate tier we're super happy with. It's our highest price tier. It's the fastest-growing tier in the company. And it's priced higher than our competitors' highest price here by multiples. And as Sid alluded to, the payback period for Ultimate is about 6 months. And in 3 years, the ROI is over 400%. And so in the deal review meetings that [ setting on in ] weekly, we do hear that some people say, Ultimate used to be 5x more expensive, now it's about 3x more expensive. And so it will just go to Ultimate versus Premium. But I think really what drives people Ultimate is the advanced security features, the compliance, the governance. And that's why we're seeing Ultimate grow greater than 50% of our bookings this quarter and last quarter and today represents about 44% of our total ARR.

On Ultimate, I said, you've described Ultimate as including the capabilities that really differentiates you from the competitors. Maybe you can sort of walk through why the ultimate tier separates you from the pack, if you will.

Yes. So Ultimate includes all of the security you would need in a DevSecOps pipeline. And that's not just static scanning, but we're the only platform with dynamic scanning, with first testing, with API security. And that allows customers to consolidate on a platform. It also makes it easier to shift security left, do it earlier, to enforce it with GitLab, you can say, hey, this project needs to comply with that and GitLab approve that the project complies for you. We're the only platform that can do that. Apart from that, we're the only platform that has integrated enterprise agile portfolio planning. We talked about the advantages of combining planning with the rest of the software delivery process. So that's the features. The #1 reason it's selling is security and compliance. The number two reason is that integration of planning.

Yes. And so it kind of opens up the conversation to the security opportunity. And to kick off that part of the conversation, there are a number of companies, not even like traditional security companies. There are a number of companies coming from dev tool world, from infra world to move into security, if you will. And so I guess the question is what gives the license for GitLab to be a credible provider of security solutions versus Datadog is trying to do this? You have private companies trying to do it. There's a whole set of incumbent point solutions that are -- that want to provide security scanning capabilities, for example, in the [ DevOps ]. So what gives GitLab the license to win in security?

Yes. I think that the breadth of it, the depth of it, we've invested heavily in the security features over a very, very long time. The only ones who can kind of do compliance automate it, the alternative is building something like that yourself. We regularly replace point solutions like Checkmarx, Nik, Veracode, Black Duck, Synopsys, and these customers are able to move to just GitLab. Super important also note is integrated into the platform and so you get the entire platform and you buy the product. And so the payback is really quick, the ability that you don't have to do all the integrations, and so there's a lot of benefits to of having it within GitLab.

Talk a little bit about the -- well, first, who owns the purchasing decision? That's the debate kind of in the market, right? You may be selling to DevOps users, SREs or developers. But is it the CISO or the SOC team that owns the budget. And to the extent that it's more on the security side, what has GitLab been doing to build relationships, both with these security buyers, but maybe with -- integrating with the security channel and the broader ecosystem overall.

Yes. So typically, software is so essential to companies. They are becoming software companies. So yes, there's multiple C-level execs, and we talk with both like the CTO, who's responsible for engineering, the CIO who wants a scalable, secure platform that complies with all their internal processes and the CISO, the person who has to do security. And if we talk to the CISO, they say, "Look, I bought all the best point solutions in the market. I just can't get people to use them." With GitLab, people are able to move and then get all those scans done. At T-Mobile, they move 25,000 software projects to GitLab in 2 months. And after that, they were running hundreds of thousands of security scans because it made it so much easier.

In terms of -- one of the teams on the earnings call has been sort of GitLab enterprise, but kind of dovetails with the security, but also the compliance and the sort of the rigorous requirements that you guys are able to meet. So GitLab dedicated, talk about why you released this offering to market and what has been the adoption to date?

So the majority of our revenue is still coming from customers who have a self-managed installation. Typically, they run that in the cloud. But what's typically important to them is that they're the only customer in that installation. That way, they're more secured. We GitLab dedicated. We can offer them the best of both worlds. Still single tenant, but it's software as a service. So no maintenance for them. We make sure it's reliable. We make sure it's fast. We make sure it's always up to date. We can get the best of the both worlds. And in Q4, Southwest Airlines became a dedicated customer.

Brian, could you frame out the uplift potential. So when a customer like Southwest Airlines, as had mentioned, they have a self-managed deployment and they go to GitLab dedicated. What does that ARR uplift look like?

Yes. So basically, the way we price Dedicated is it's -- you have to be an Ultimate and you have to be a certain number of licenses and then there's the infrastructure charges on top of it. I think it's less about the individual customer and more about the addressable market for more customers with a more complex environments that would like this to get up and running faster on GitLab.

So when we think about that as just dovetail off of your response. I should think about the self-managed base and potential penetration of GitLab enterprise within that basis, maybe as kind of the first step before they go full haul cloud. Is that the best way to think about it?

I think it's a way to -- we're the only DevOps platform who offers single tenant sound. And I think the customers will get more value. And if they get more value because they're always up to date and it's always reliable, they tend to -- it tends to grow. They tend to get more people involved. They tend to add planning to it and everything else. So it's not so much that dedicated itself makes money. It's so much of a better experience for the customer and the users that we tend to grow seat count within those customers.

Yes, makes total sense. Let's talk a little bit about the underlying tech. I mean, so you guys are delivering a lot of different product capabilities. Underpinning all of these products is a singular data store. Can you talk about the architectural decision to have one massive back end for the platform? What types of data gets housed in the GitLab data store and what kind of intelligence and future capabilities will this approach unlock?

Yes. There's other providers who call it a platform, but it's kind of multiple applications. And then you're still kind of switching between apps. I think us and GitHub are the only true platforms today, where it's a single application, single code-based single data store, and the big advantage of that is you get a faster cycle time. You don't have to switch applications. You can do very intelligent things with AI, for example. We wouldn't have been able to have so many AI functions the most in the market, if it was across different applications. So we're very proud. We have the broadest platform because the more point solutions our customers can replace, the more they save on licensing, the more they save on the integration costs, the better they can up their compliance, the more efficient they get.

Yes. It's really interesting point, because that's what you're saying, it translates to product velocity, right, in terms of the breadth and different capabilities and your ability to go from more 1 product...

Exactly. Like -- companies are becoming software companies, typically the speed at which they can move is the speed at which they can improve their software. You can go 7x faster, that's a massive advantage.

Makes total sense. Let's talk a little bit about go-to-market. And to kick off that part of the conversation, we obviously saw a bit of a boom coming out of the pandemic calendar -- late parts of calendar 2020, 2021, early 2022. What parts of that environment, that sales playbook that worked back then, what part of that is still valid today? And what parts are that we have to sort of chunk out because that was just a really special time in the market and really frothy time to the market where budgets were plentiful.

I think with us, there's no revolution. It's evolution. We have a new CRO, Chris. He hit the ground running. One thing that he's doing is he's focusing even more on the largest customers -- because the bigger companies are, the more complexities to keep all these point solutions up in the air, the bigger the benefit of moving to a platform and we want to make sure each and every one is successful, not just in the purchase, but implementing it and getting realizing all the value.

Makes a lot of sense. Talk a little bit about Chris Weber, your new CRO. What impact has he had on the sales organization, maybe from an operational perspective, building muscles. You talked about focusing on larger customers. But as sort of the quarter-to-quarter execution, how -- what role is it going to play in GitLab executing against its targets?

Yes. He inherited a great organization. He's doing an amazing job. I think the numbers we put up in Q4 are an example of that. The biggest focus have been in order to be a great kind of technical thought partner to our customers and to make sure that when they vouch for GitLab and put their job on the line to say, look, we're going to get rid of these 10 point solutions. We want to make sure they're successful each and every time. We're a pure-play company. This is what we do. We do one thing and we want to do it very well each and every time.

Let me go back to the security topic and the go-to-market behind that. Sometimes when we're trying to build relationships with these security purchase decision makers, there's an overlay specialist security sales force who may have had experience with those customers and those CISOs and the soft teams, if you will. Is that an approach that you guys have considered? Is there going to be an overlay specialist security sales force at GitLab? Or is it going to be all salespeople selling the entire platform?

Yes, all of our salespeople have to be able to talk both with the engineering leader, with the IT leaders, and with the CISOs. It doesn't mean we don't have specialists in GitLab. And only the best person to talk to your customers is Josh, our Chief Information Security Officer. He's part of our E-Group, he's on an equal level and he has a lot of customer empathy and talking to lots of CISOs at our customers.

Probably from an investment perspective, is there a way to think about how much like a security sales motion, is that require more investment versus what we've seen over the past couple of years when you think about allocating investments?

I think it's really -- that's hard to say. I think if you look at the money companies that are prepared to spend, typically, it's higher in security. You can defer from us -- Ultimate more than half of bookings, 44% of revenue, right, growing. And the main reason to buy it is security and compliance. So we're becoming more and more a security company, even though we also have agile planning and deployment and everything else, but security is such a compelling reason that that's a bigger and bigger part.

I have one more question on the product and the platform, and then we'll give the audience to see if they have any questions for the GitLab management team. The simple question is this, like how does the team decide just given that you have so much product velocity? How do you decide what features reside in a premium subscription versus ultimate like how do you guys think through that decision?

Yes, we have a model for that. We call buyer-based open core. And basically, it's -- if an executive really wants it, it's an Ultimate. And for example, an executive wants to prove compliance to the [ auditors ]. If a director wants it, we put it in Premium. If an individual contributor wants it because it makes their day-to-day faster, we put it in the free open source version to drive adoption.

Yes. That's really interesting to think about that. Let's go to the audience. If you just raise your hand and wait for the microphone. We'll kick off the Q&A. Here in the front.

I wanted to ask about -- you talked about that developers. Only 25% of their time are writing code and the other are working on other stuff. We're seeing new models coming with huge context windows of 1 million tokens. We're seeing people uploading an entire code base and the LLM can analyze it. I was wondering where do you see the new initiatives in automated testing? And what is the GitLab's plan to automate the other area of the stack?

Yes. These models are getting better. They will be able to get more contracts. I think we're in a great position as a platform that host people's code to add that contact and to give it to the model when relevant. And GitLab already has a feature to help you write tests. And that's going to -- those things will continue to grow, get better, get more automated, require less human handholding.

I see we had a question up front.

I got 2 questions. One is on the potential -- growth potential of the company. So you talked about that we are consolidating many points. And then can you walk us through the potential of the TAM because a lot of the holders are not paying yet, but paying like, I don't know, $20, $30 or even $50 a month. It's not a big expenses, right? But -- so how should we think about that in terms of the user and ARPU? And my second question will be on competition with GitHub, right? So let's say you go to the CISO, then how do you convince them to between you GitLab and GitHub?

Thank you. The total market for kind of DevOps tools is estimated to be $40 billion revenue a year market, and it's growing. So it's a big market, that's in GitHub together or less than 5% of the market. The whole name of the game is to our place to point solutions faster and GitLab today can replace more point solutions than anybody else. Another way to kind of size the market is to look at, for example, a go-to-market. We have a go-to-market organization. They're paying for sales force, they're paying for this, they're paying for that. They're spending a lot more per person than it's currently being spent in the software development market, even though those people are being peak typically as much money. And I think we can have as big as an impact with software there. So we feel really good about the size of the market. We're talking to a CISO and it compares to GitHub. They see we can do a lot of things that they need and that GitHub can do. GitHub doesn't have dynamic scanning. GitHub doesn't have API security. GitHub doesn't have us testing. Even better we can prove that they're compliant. With GitLab, they can tag every project, no to security framework, they -- what that security framework consists of and prove that. The auditors can walk in, point to any environment and they have the reports ready. We're unique in that capability. Oh, you asked how we would talk to the CISO? If I were to talk to a CIO, I bring different points up. I would GitLab's installation. It's more secure, it skills better, it gives them more ways to customize. Like talk to an engineering leader, I talk about how it's a better user experience because we replace more point solutions. And you can finally tell the developers they can [ bifurcate there ] and they all cheer for him or her. So it's a different talk track depending on who's listening.

Go back to the audience for Q&A, but I wanted to pick up on his point. When the generative AI, like kind of hype cycle hit in full force at spring of last year. There was an initial view that when things like GitHub copilots that would be disruptive to companies like -- companies like GitLab. And since then, we've seen a number of companies released their code adviser, copilots, you guys have as well. And so going forward, how does the customer think about one? First part of the question. What AI is going to do to other parts of the life cycle, so beyond code advisers? And what's going to be the basis of why I choose one platform like a GitLab versus a GitHub and there's even other players that are trying to do some other things. Like what's -- how has AI changed the basis of competition is the spirit of the question?

Yes. Great question. Three points I'll make. First one is you need a good code suggestions functionality. We went generally available with code suggestions in December, it is a competitive offering. The second thing is the context windows are increasing, like the AI can get smarter, if you give it the relevant information. The code hosting platforms, GitHub, GitLab, have a natural advantage in making DIs give smarter suggestions because it's easier for us to give that contract. Third thing is people are rising up. AI is not just writing more code. It's helping the developers with everything else. It's helping their security, the operations, the planning people. And they're starting to look at, okay, who can add the most value over the whole software development life cycle. Recent Omdia report, we met 37 of the 42 use case, more than any other DevOps platform, and that's starting to become a bigger and bigger thing. So the shine of the headlining feature is kind of -- yes, and we don't [ scatter ] that, and now it's about who can add the most value throughout the life cycle.

Makes perfect sense. Let's go back to the audience and see if there's any questions for the management team. Just raise your hand, and we'll get the mic to you. No questions. All right. So let's talk about enterprise agile planning, right? It's been a theme in a couple of your last earnings calls. The decision to move into planning has obviously been a huge area of strength for Atlassian and this Jira offering. I'd love to get a sense of like the scale or the momentum behind customers sort of rethinking the planning piece and bringing it into a platform like GitLab.

The great thing about planning is that there's a lot of new people involved. So we had a customer. We talked about it during earnings. They moved to GitLab thousands of people, but they bought additionally thousands of extra licenses for their salespeople. So for us, these are kind of net new people getting involved. We needed a new SKU for that. So we have enterprise our job planning at $15 per user per month paid annually. And that's because these people are not using any of the other features of GitLab. If the agile planning is included and Ultimate, if you're doing destack open, you need planning, it's part of that. But these new users coming in is very exciting. We're very, very, very early in this movement. Most customers today are in Jira, and we're starting to see the first customers. UBS was kind of the first mover on that. And now we're starting to see all the other ones coming in. But it's early. It's not going to have a giant effect on our revenue for this year, but we see the potential over time to -- it's the best thing for all of our customers to bring planning into the platform. So we see a great long-term potential there.

Great. Maybe let's wrap up on just the investments required to scale AI capabilities in the AI portfolio. Brian, maybe this is a question for you. If not, feel free to take it. But as you look at your overall investment philosophy as early to AI, to what extent will your AI investments require significant investments in GPU capacity? And to what extent will that impact your COGS?

So when we look at our investments in R&D in general, one thing that's unique about GitLab versus a lot of other companies, we have single engineering groups, we do monthly software releases and so we iterate a lot. And so it's not like we're trying to make a monolithic system, but a whole bunch of people on it, launch it 6 to 9 months later. And so we get a lot more done a lot quicker than other companies. And so the investments that it takes we've prioritized because we're working on the full breadth of the platform are in security compliance around AI, and that's built into the numbers that we've given out.

Awesome. With that, we'll leave it there. Thank you so much, Sid and Brian, for coming to the TMT conference and give us an update on the GitLab story. We really appreciate it.

Thanks for having us.