Home / Transcripts / GitLab Inc. (GTLB) · March 3, 2026

GitLab Inc. (GTLB) Earnings Call Transcript

March 3, 2026

NASDAQ US Information Technology Software earnings 62 min

Earnings Call Speaker Segments

Moderator operator
#1

Good afternoon, and thanks for being with us on the call today as we discuss GitLab's financial results for the fourth quarter and fiscal year 2026. Joining me are our CEO, Bill Staples, and our CFO, Jessica Ross. Over the next while, we'll walk through a business overview, share our commentary on the fourth quarter and the full year, and offer guidance for the first quarter along with fiscal year 2027. Ahead of getting started, let me handle the safe harbor statement. I'd point you toward the cautionary statement about forward-looking statements, which appears on page two of our presentation and also in the earnings release we put out earlier today; you can find both in the Investor Relations section of our website. Within the presentation and the earnings release, there's a discussion of particular risks, uncertainties, assumptions, and other factors that might cause our results to differ from what's expressed in any forward-looking statements under the meaning of the Private Securities Litigation Reform Act. As usual, this language governs the content of today's call and presentation. Also, on today's call we'll be discussing certain non-GAAP financial measures; such non-GAAP financial measures leave out certain unusual or non-recurring items that management believes affect how comparable the periods referenced are. For more information on these non-GAAP financial measures, and for the reconciliations to the most directly comparable GAAP measure, please refer to our earnings release and presentation materials. Now I'll hand things over to Bill. Bill?

Bill Staples executive
#2

Thanks, Yao. Everyone, good afternoon. For GitLab, fiscal 2026 proved to be a major year. Our ARR crossed the $1 billion mark. Free cash flow came to $220 million, which is a gain of more than 80%, alongside margin expansion of almost 7 percentage points compared with the prior year. Both FY 2026 and Q4 set records for us in absolute net new ARR, for a full year and a quarter respectively, and we plan to keep building on that momentum. The effort our team is putting in makes me genuinely proud. With the technological and structural edges we hold, plus a TAM that keeps growing, we think there's room to perform even better. I'm confident the right people are here to go after this opportunity, and I'll use the remainder of my comments to detail our plan. Jessica will take you through the financials. Let me speak plainly about our FY 2027 outlook; our revenue growth guidance does not satisfy us. As with lots of businesses that get to a billion dollars in revenue, our attention has increasingly turned toward scaling growth. We've pinpointed five particular strategies offering the biggest chance to boost growth at scale during FY 2027. Those five are: first, getting first orders accelerating again to drive expansion over the long term. Second, growing sales capacity through committed leadership and funding. Third, broadening product packaging to open up fresh monetization avenues. Fourth, reaching price-conscious buyers with more value and wider coverage. Fifth, staying the course on an AI strategy that matches the strengths of our core platform. Work on all five is already underway. FY 2027 comes down to execution and confirming our hypothesis through results. Let me go through each; during FY 2026, we ended a lengthy stretch of first order deceleration. That matters enormously, since our customers typically start small yet keep expanding, a pattern that has persisted for more than 10 years. Sales-led first orders started accelerating again in Q2 FY26, immediately after Ian came aboard. On the product front, Manav has brought new energy to product-led growth. First order logos turned upward in October, and improvement has continued for four straight months. Looking at FY27, we now have a clear route to lasting acceleration in first orders, powered by ongoing sign-up momentum, additional product-led on-ramps, and a committed first order sales team featuring a new global leader, four regional leads already appointed, and fast-paced hiring; indeed, that team has closed its first deals in Q1 already. As evidence, during this quarter we landed a landmark deal with one of the semiconductor sector's most strategic companies, a cornerstone supplier within the AI super cycle. Following a competitive evaluation pitting us against incumbent tooling and AI-powered options, they picked GitLab Premium and Duo Enterprise for more than 5,000 users, which affirms our unified platform and AI capabilities. We started adding headcount during FY 2026, and as FY 2027 begins we carry more capacity than at any prior point; we have a route toward even stronger ramped capacity starting in Q3. With the FY 2027 kickoff, we've redesigned territories to serve every segment better and to reinforce enablement. On the innovation front, GitLab has spent a long time consistently providing value, with 172 straight months of fresh releases. Clients who consolidated their repositories and CI pipelines onto GitLab continue to look for ways to broaden their usage with us, but they tell us our pricing model feels too one-size-fits-all. Over the course of FY 2027, we plan to introduce a number of new monetization opportunities on a quarterly basis — among them built-in artifact management, software supply chain security, and integrated secrets management, along with other additions. These will be optional, à la carte products that give customers of both GitLab Premium and GitLab Ultimate a middle-ground path, since they've been asking for ways to buy extra value at an incremental cost. Because most of these are expected to arrive spread across the year, we see only a modest contribution in FY 2027, with a more meaningful effect in FY 2028 and afterward. A few years ago, we raised GitLab Premium's price by 50%, and that coincided with increased experimentation around AI code as well as SaaS budgets that were roughly flat. Meanwhile, as we moved upmarket, we reduced technical resources serving the lower end of the market. Together, those things have slowed GitLab Premium's growth, particularly with customers who are price-sensitive — a group we estimate makes up roughly 20% of our ARR, which includes the SMB softness we've been discussing lately. We're tackling this in a number of ways; starting in mid-January, we have an AI product, plus a platform and market, that helps accelerate the whole software life cycle. We're bundling attractive GitLab DAP promotional credits for Premium and Ultimate users to boost the value they perceive. We've also tweaked coverage models for this cohort, and we're putting investment into onboarding, adoption, and self-service experiences that will assist every customer in obtaining value more quickly. GitLab occupies the center of how enterprises construct and ship software. In January, we introduced GitLab Duo Agent Platform and reframed GitLab for the AI era. Our intelligent orchestration platform enables users to place AI agents throughout the software life cycle, drawing on the same context, permissions, and security model that they currently have in place. This platform stands on three core pillars. Workflows gives teams and AI agents one shared space to collaborate on tasks that cover the entire software life cycle. Context delivers rich semantic access to complete SDLC data, which leads to higher quality and more efficient outcomes. Guardrails. With GitLab, deployment can happen anywhere, and security plus compliance are built right into your software factory. GitLab sits exactly where AI systems deliver the most value: the point of execution. We pull together the context that's missing and act where the code actually lives — where it gets merged, built, deployed, where compliance is enforced, and where corrective actions stop downstream bugs, technical debt, and security issues before they spread. Even before AI, our platform cut down on friction for developers. At this point, it can deliver step function productivity gains by making things easier for agents and the people overseeing them. Alongside our seat model, Duo Agent Platform brings in usage-based pricing. Agent work is what customers pay for, and each engineer can hand off tasks to several agents running in parallel. As agents take over more of the software life cycle, our revenue rises in step with the value delivered. Consider an airline customer of ours whose technology organization numbers 3,000 people. They are using Duo Agent Platform to automatically handle vulnerability fixes, dependency upgrades, and cloud migration tasks. Nowadays, roughly 90% of their component updates occur with no human involvement, freeing developers to focus on building features for customers. Our roadmap is ambitious, and we intend to ship something valuable each and every month, with a focused go-to-market push to speed up adoption. Just to remind everyone, close to 70% of revenue is generated by self-managed customers, and those customers need to move to release 18.8 or higher, and what we usually observe is that it takes two quarters before more than 50% of them move to the new release. Wherever we can, we're putting money in together with our partners to speed those upgrades along. FY 2027 is focused on turning pilots into production deployments, not on bringing in significant revenue. Once metrics become material, we'll share them. A fundamental shift is happening in the software development market. The pace of AI development just keeps accelerating. That's causing the amount of code being produced to balloon, along with the complexity of getting it delivered—and when something goes wrong now, the fallout is more severe than at any point before. You can't afford to set aside security, compliance, and governance. Those things determine whether you survive. GitLab's entire design was a response to exactly this set of conditions. What I've laid out—restoring our ability to go to market, opening up fresh ways to generate revenue, and putting GitLab right at the heart of Agentic AI—these are not three distinct efforts. They form a single, unified strategy aimed at winning a market whose momentum favors us. The numbers back this up. During Q4, we brought in more $1 million customers than in any other quarter of GitLab's history.

Bill Staples executive
#3

On the retention side, we're seeing numbers that track with what we've historically experienced, and churn has never been lower over the past four years. Ultimate makes up 56% of ARR at this point and was part of nine out of our top 10 deals. Growth in Ultimate projects that include security scanning exceeds 60% compared to a year ago, and we're seeing close to 30% more security projects for each seat. Indeed runs the top job site globally. Their GitLab journey started back in 2015 with source control, then they moved up to Premium in 2020 to help with CI/CD adoption, and in 2024 they made the jump to Ultimate for advanced security, compliance, and governance across thousands of GitLab users, which delivered an 80% boost in pipelines while bringing infrastructure costs down. This quarter they're taking their strategic partnership further by moving to GitLab Dedicated as they continue modernizing their infrastructure. The Mercedes-Benz expansion we saw this quarter is another good example of the compounding growth potential we have. Vehicles being built today have more software code in them than fighter jets do, which pushes companies like Mercedes to bring on thousands of engineers. Our connection with them started years back around source code management. These days, GitLab acts as the central platform that powers their software-defined vehicle transformation, serving thousands of developers in different regions. It makes sense that investor uncertainty is elevated right now. Once every developer can tap into the same models, code generation turns into a commodity. The real bottleneck moves to whatever comes after the code. Reviews, security, pipelines, compliance, deployment; that's exactly the space we occupy, and that positioning becomes increasingly difficult to copy as AI spreads. A number of our customers are already dealing with decades of technical debt, thousands of repositories, and compliance obligations connected to policies that were written years back. All of that context sits within GitLab. History, ownership, risk, intent. Everything is being indexed and linked throughout the software life cycle. When it comes to autonomous agents, context determines whether an action is useful or potentially catastrophic. Each commit, each scan, each deployment enriches our graph and improves our agents' accuracy. The longer someone stays on GitLab, the more intelligent the platform becomes; that's a mode that expands over time; with GitLab Duo Agent Platform, we're more than just a tool that agents utilize, we're the environment they operate in, the orchestration layer that controls what they do, the sequence they do it in, and the guardrails they stay within. We possess the ingredients needed for a generational company, a market that keeps growing, trusted distribution at scale, deep customer relationships, and platform capabilities built up over years. Our financial position is strong, with roughly $1.3 billion in cash and investments, and we're generating free cash flow in a sustainable way. I'm happy to report that our board has approved GitLab's first share repurchase program, totaling $400 million, which shows confidence in our fundamentals and the growth plan ahead. We think GitLab shares are attractively valued and we remain dedicated to disciplined capital allocation. FY 27 is about proving that this foundation can deliver value to customers, momentum through consistent performance, and progress quarter-by-quarter. With that, I'd like to hand things over to Jessica to walk through the financial results.

Jessica Ross executive
#4

Thanks, Bill. And thanks to all of you for being with us today. As GitLab's new CFO, this marks my first earnings call, and I'm glad to be part of it. The reason I came aboard is that I saw a remarkable company sitting right at the heart of an industry shift unlike anything we've seen. Being able to influence the way AI reshapes software development via intelligent orchestration is an opportunity I find very compelling. During these initial weeks, I've been struck by how devoted the customers and teammates are, how deep the platform's technical capabilities run, and the quality of the leadership group Bill has put together. My priority is instilling the financial discipline and operational rigor needed for our next phase of growth. I'm eager to get acquainted with many of you over the weeks ahead. I'll begin with the results for the full year and the fourth quarter, then walk through how we think about capital allocation along with our FY 2027 guidance. Fiscal 2026 proved to be a robust year. Revenue climbed 26%, reaching $955 million. On a non-GAAP basis, operating margin came in at 17%, roughly 680 basis points higher than the prior year. Adjusted free cash flow rose 83%, hitting $220 million, alongside more than 7 points of margin expansion; at present we count 10,682 customers whose ARR is at least $5,000, and they supply over 95% of total ARR. The group spending $100,000 or more expanded 18% versus a year ago, reaching 1,456 customers, which accounts for slightly above 75% of ARR. As Bill noted, Q4 brought in the highest count of $1 million-plus customers ever for GitLab, now exceeding 155, a 26% increase year-over-year. Now let me turn to what happened in the fourth quarter. Q4 revenue came to $260 million, a 23% rise year-over-year and 3.5 points ahead of guidance. Non-GAAP operating margin hit 20.5%, beating guidance by 5 points. Part of the revenue outperformance stemmed from roughly $3 million in one-time items tied to favorable foreign exchange and JiHu's performance. First order bookings looked healthy, especially in Asia-Pacific. Enterprise win rates rose compared with the previous quarter, and sales cycles stayed steady. Performance in the U.S. was softer, though. More broadly, a handful of large deals slipped away as customers dealt with budget constraints and industry challenges. We likewise saw just a partial rebound in the public sector after the government reopened, plus ongoing softness in the price-sensitive cohort Bill referenced earlier. Dollar-based net retention stood at 118%. Gross retention continues to sit comfortably above 90%, in line with historical patterns. Our biggest customers keep expanding, even as pressure in the mid-market and SMB segments dragged on net retention. Total RPO increased 20% year-over-year, reaching $1.1 billion. Current RPO rose 24%, to $719.4 million. Non-GAAP gross margin was 89%. SaaS accounts for roughly 32% of total revenue now and expanded 38% year-over-year, fueled by ongoing momentum in GitLab Dedicated and GitLab Duo. Q4 non-GAAP operating income totaled $53.4 million, an operating margin of 20.5%, up roughly 280 basis points year-over-year. Adjusted free cash flow for Q4 was $41.8 million, a 16% margin. We closed the quarter holding $1.3 billion in cash and investments. Turning to JiHu, Q4 non-GAAP expenses came to $3.9 million, versus $3.2 million a year earlier. Deconsolidating JiHu continues to be what we're aiming for, but there's no way for us to forecast how probable that is or when it might happen. Before we get to guidance, I want to quickly walk through how we think about allocating capital. Number one is putting money into growth. The first dollars go toward the investments with the highest returns — the ones that speed up our product roadmap and make our go-to-market motion stronger. That means expanding sales capacity, standing up our first order team, moving faster on Duo Agent Platform, and pushing security innovation deeper; we're shifting resources toward the initiatives with the best returns while keeping growth, profitability, and cash generation in balance. Number two is keeping the balance sheet resilient; we're holding a strong liquidity position. Thanks to sustainable free cash flow along with roughly $1.3 billion in cash equivalents and short-term investments, we can put money into ourselves and into inorganic growth across cycles with no constraints. Number three is buying back shares. GitLab's board has approved our first ever $400 million repurchase program, which shows confidence in our fundamentals and a disciplined stance on capital allocation. We see buybacks as a meaningful lever for creating shareholder value and managing dilution, especially during stretches when the share price is dislocated. Now on to our FY 27 guidance. This guidance is our clearest read on the business given how things stand operationally right now. Let me lay out the key assumptions ahead of the numbers. First, ratable model dynamics; under a ratable model, revenue captures the cumulative impact of bookings activity across multiple earlier periods. As Bill said before, we're not content to just sustain historical growth rates, and we're pursuing all five growth opportunities he laid out. A meaningful chunk of our FY 27 guide reflects the fact that bookings growth hasn't scaled alongside revenue over the last three years, and that reality works through arithmetically. Non-recurring FY 2026 tailwinds. FY 2026 got a lift from a number of items that we aren't building into our FY 2027 guidance. Taken together, that's roughly 300 basis points of growth. In rough order of size, these are the premium price increase from three years back, favorable FX movements, and certain provisions in some customer contracts. Third, segment caution; we're expecting the public sector to do better in FY 2027, but we aren't modeling a bounce back. For the price-sensitive cohort, which is about 20% of ARR, we expect continued pressure based on what we saw in Q4. And finally, prudent assumptions on newer growth drivers; we're assuming minimal revenue contribution from GitLab Duo Agent Platform in FY 2027; it launched seven weeks ago and converting pilots into production deployments takes time. On top of that, roughly 70% of our revenue comes from self-managed customers, which means adoption will follow a measured curve. With all that as context, for Q1 FY27 we expect total revenue between $253 million and $255 million, which is roughly 18% to 19% growth year-over-year. We expect non-GAAP operating income of $32 million to $34 million; we expect non-GAAP net income per share of $0.20 to $0.21, assuming 173 million weighted average diluted shares outstanding. Keep in mind that Q4 includes 3 additional days compared with Q1, which sets up a sequential headwind. For the full year FY27, our expectation is total revenue between $1.099 billion and $1.118 billion, which works out to roughly 15% to 17% growth year-over-year. Non-GAAP operating income is expected to land between $129 million and $137 million. We're also projecting non-GAAP net income per share in the range of $0.76 to $0.80, based on an assumption of 175 million weighted average diluted shares outstanding. A few more modeling details. We're assuming year-on-year growth rates stay stable from Q2 through Q4. Gross margin for the full year should come in at 85% to 87%, which is lower than the 89% we saw in FY 2026, and that reflects a larger mix of SaaS, Dedicated, and GitLab Duo Agent Platform, each of which has a different cost structure. When modeling FY 2027, we're forecasting roughly $15 million in expenses tied to JiHu, versus $13 million a year ago. We're operating from a position of strength. Our TAM keeps expanding. Customer retention is still best in class. Our biggest customers keep growing, and first orders are back to growth. We're building new multi-year growth drivers through GitLab Duo Agent Platform and hybrid pricing. Our FY27 guidance reflects where we stand right now, early in a transformation with clear priorities: scaling sales capacity, stabilizing net retention, addressing the price sensitive cohort, and moving DAP pilots into production. As we make progress, we'll keep you updated. Thanks for joining us. I'll now hand the call over to Yao to moderate Q&A.

Koji Ikeda analyst
#5

Yeah, appreciate you fielding this one, Jessica — good to finally connect with you on the call. My question is about security. Security plays a major role in driving ultimate upsells, and presumably in agentic duo security adoption down the road as well. So given all the buzz around Claude Code Security, plus the possibility that other foundational model vendors could roll out their own code security offerings, how are we supposed to view the differentiation GitLab delivers through its security portfolio? What makes it keep fueling ultimate upsells and Agentic usage going forward? Thank you.

Bill Staples executive
#6

Thanks, Koji. That question comes up with me frequently; the heart of it is really the gap between suggestions and certification. Claude Code Security assists developers in producing better code while they're authoring it; that's genuinely valuable. A tool offering secure code suggestions at authoring time cannot also be the tool that certifies something is production-ready. That's the role GitLab plays. GitLab serves as that independent system addressing a separate question. It addresses whether this project, this source code, is ready to ship. During authoring time, a developer is free to disregard Claude's recommendations. In fact, Anthropic's own page states that developers always make the call. When it comes to GitLab, they cannot bypass the execution policy for the pipeline. The merge re-request approval rules can't be ignored either. Companies ship what is secure and meets their engineering standards. GitLab is the platform they depend on to uphold those standards. Truly, these two are complementary. Claude enhances source code at authoring time, while GitLab governs whether the software is permitted to ship.

Matt Hedberg analyst
#7

Thanks a lot, and thanks for fielding my question. Yao, appreciate it, and hello everyone. Bill, at the top of this call you noted that you aren't pleased with where the growth targets stand for this year. I definitely value the five initiatives you laid out. It sounds like there's plenty there that could drive growth. Thinking through a few of the points Jessica raised regarding the fiscal 27 guidance, my question is this: what's the right way to think about when acceleration arrives, and even the route back to that 20% figure, or stronger growth, which I'm certain is what you're aiming for?

Bill Staples executive
#8

Sure, that's a reasonable question. When I look at those five growth initiatives and consider which one delivers the largest near-term effect on FY27, the answer really begins with what we're putting into go-to-market. More capacity lets us serve our current customers better while also bringing in new logos faster. As I said earlier, we're heading into this fiscal year with more capacity than we've ever had, and beginning near Q3 we anticipate a step function rise in ramp capacity. That's my framing for GitLab in FY27, though if we zoom out and consider growth over the long haul, let's keep in mind that we just posted the biggest new net ARR both for a year and for a quarter in our history. Our core business is genuinely strong. Gross retention stands at its highest point across the past four years. Each customer cohort since inception keeps expanding. Win rates haven't moved. Engagement keeps climbing. This business has been slowing owing to bookings patterns and lapping mechanics across the past three years. It isn't becoming less relevant. Actually, its relevance keeps building momentum in the age of AI. To tackle the value capture equation, that's the reason we're going after several new strategies. Alongside added capacity to chase the TAM, we're bringing in those new SKUs so customers can opt into extra adjacent value. That's also why we've now rolled out Duo Agent Platform featuring a fresh hybrid pricing model, letting customers capture value and automate tasks across the full lifecycle. We're able to bill according to work performed and value provided, rather than only by seats. It's likewise why we're revising our coverage models and putting money into included DAP credits for customers within that price-sensitive cohort, boosting their value equation too, along with their stickiness and growth. Over the long run, I'm confident this company holds everything required to become a high-growth generational company, and executing on that begins right here in FY27.

Matt Hedberg analyst
#9

Appreciate it, Bill.

Rob Owens analyst
#10

Excellent. Many thanks for fielding my question. Bill, I'd like to expand on something you said regarding gross retention sitting at record highs, and yet net retention is declining. I'm aware you brought up, you know, certain softness within the mid-market. Perhaps to start, could you break down that NRR figure for us a bit and explain what's behind it overall. You know, clearly there's a bit of worry regarding seats out there and possible seat expansion among customers. Give us insight into that dynamic. Then I suppose, second, since you're steering toward total revenue below the current level of your retention rate — your expansion rate, rather — you know, perhaps some boundaries on where NRR might land across the next year. Thanks.

Bill Staples executive
#11

Yeah, I'll handle the first portion, and then perhaps Jessica can address a little about where.

Jessica Ross executive
#12

Yes.

Bill Staples executive
#13

Regarding FY 27, here's how to approach it. Dollar-based net retention isn't a metric we provide guidance on or center our attention around. It's something we view as an outcome generated by the business. When examining how DBNR is distributed among segments, enterprise looks quite robust to us. The 100K cohort, which I mentioned earlier, expanded 18% compared with the prior year, and it accounts for 75% of our ARR. Growth was even quicker for the $1 million cohort, at 26% year-over-year. Expansion among our biggest customers keeps happening, and that indicator is genuinely powerful and steady. Where we feel strain is mostly within the cohort that's sensitive to price. We put that at roughly 20% of ARR, encompassing the 8% portion of the business discussed earlier, which sits in SMB, plus certain areas of mid-market and premium. In FY 27 we're tackling this by now bundling DAP credits into each Premium seat; we've reworked coverage models so they connect better with GitLab, adding technical services to speed up, you know, adoption and realization of value. We're putting money into improved time-to-value experiences. Jessica, would you like to say a bit about the figure itself, and where we anticipate it heading next?

Jessica Ross executive
#14

Right. Like you mentioned, and like Bill mentioned, we don’t provide guidance on that figure, but I want to emphasize that for GitLab, this is the year of stabilization. So I’d expect DBNR to decline a bit before leveling off.

Sanjit Singh analyst
#15

Yes. Yes. Are you able to hear me?

Moderator operator
#16

Yes, we can.

Sanjit Singh analyst
#17

Great. Thanks for taking my questions, and Jessica, congrats on the new position. Bill, my question is basically about pricing. How did you, you know, how did you land on that pricing formula for DAP, for our Duo agents? The way I'm thinking about the context is that coding agents, the ones that are popular, many of them still use seat-based pricing, correct? Be it $20 per month or more. This one is more — DAP appears to be priced more like a consumption-based model. How did you get to that pricing mechanism for DAP, and then looking ahead, how do you think about capturing additional value within the GitLab platform?

Bill Staples executive
#18

Okay, I'll begin by addressing what you noticed about rivals appearing to rely more on pricing per seat. In reality, that isn't completely accurate; an entry cost tied to seats may exist, yet afterward they either limit usage or bill you for exceeding limits, and that approach is genuinely less efficient compared with what GitLab provides, since we give customers the option to begin with credits that are included. With each Premium seat, there's $12 in credits. With each Ultimate seat, there's $24 in credits. The reason is that we aim to capture their hearts and minds during this stage of the platform; they're able to use it without needing to sign new contracts or obtain new agreements. Should they discover value, they can subsequently decide to opt in to on-demand credits. On-demand credits can be understood as essentially, you know, pay as you go. We issue a monthly bill tied to actual usage at roughly $1 a credit. What that accomplishes is establishing a demand model for our sales force, in which they observe the signal of customers using and receiving value, and they can then hold a conversation with the customer and extend additional discounts for committed credits as a monthly minimum. That propels the flywheel of ratable revenue and ARR, since those monthly minimum commitments become a subscription that we recognize ratably. This is an extremely powerful model that we believe delivers better value to customers than needing to pay on a per seat basis for throttled usage or overages. It's more efficient. I'll add just this: the competitive tools already out there are heavily subsidized by venture capitalists. My view is that won't persist indefinitely. As a matter of fact, I'd forecast that within this year you'll witness many of the enterprise tools starting to shift toward API-based charges rather than seat-based charges, which could increase bills for enterprises in the years ahead.

Karl Keirstead analyst
#19

Yes. Is my audio coming through at this point?

Moderator operator
#20

It is. Hello there, Karl.

Karl Keirstead analyst
#21

All right. Thanks for that. My apologies for the hiccup. Hi, Jessica. I'd like to dig in a bit regarding the first non-GAAP margin outlook for fiscal 27. At the top end it appears to be 12%, and that represents a 5-point slowdown versus the year you just reported. Could you break that down for us somewhat and walk us through a few of those investments? My hunch is that sales capacity, and maybe even certain DAP free credits, might be dragging on that margin projection, though I'm eager to hear what you think. And stepping back — I realize you won't provide guidance — how are you and Bill, in a conceptual sense, viewing the margin framework here at GitLab, and do you believe this year might mark the bottom, and how much does it matter to you and Bill to lift those non-GAAP EBIT margins to, let's say, 20% or more? Thanks.

Jessica Ross executive
#22

Thanks for that, Karl. So our FY27 margin guidance bakes in three separate, well-understood investments — one we actually committed to when we went public — and from there, we feel we have clear visibility into expansion. Roughly 300 basis points of the decline at the midpoint stems straight from the gross margin mix shift we've already walked through. There's the SaaS mix transition we flagged ahead of the IPO, and the rest of the compression comes from two very intentional investments. Number one, as Bill mentioned, we're rebuilding go-to-market capacity that went under-invested across several years. We're growing sales capacity, standing up our first order team, and going deeper on partner coverage. Number two, we're speeding up the Duo Agent Platform while strengthening security and innovation. Our investment priority is straightforward. In my remarks on the capital allocation framework, I laid it out: R&D comes first, then sales and marketing, then G&A. Nothing here is structural — every piece has a set timeline and a clear return as the business scales. Looking ahead, this is genuinely a year of investment, and we see it as the right call for creating long-term value, and we're going to track the payoffs very closely while giving the business the right guardrails. That discipline hasn't changed. Part of why I joined this business is that it has shown it can grow profitably and responsibly. The proof is the 1,700 BPS of margin expansion delivered across the last two years. Over the long run, we don't manage to a margin percentage; we manage to gross profit dollar growth and the durable returns that come with scaling our platform.

Karl Keirstead analyst
#23

All right. Thanks.

Shrenik Kothari analyst
#24

Yes. I appreciate you fielding my question, and Jessica, glad to have you here. I'm eager to collaborate with you. Bill, you pointed out that fiscal 27, naturally, centers mostly on the shift into production, with the monetary upside still fairly limited. I realize you addressed the schedule somewhat when answering the earlier question, in terms of growing sales and restoring the go-to-market motion. Based on a bit of initial input from customers — and there's one customer you brought up — could you share, at a really broad level, which possible bottlenecks on the customer side you've noticed up to now? Could it be confidence in these workflows powered by AI? Or is it more about approvals around governance? In your opinion, what makes it possible to effectively grow these into production settings? Many thanks.

Bill Staples executive
#25

I'll begin by discussing the initial reactions we've received from customers regarding DAP, and after that I want to remind those investing with us about what's required to turn those first impressions and trial runs into actual money coming in. So, the thing is, we're receiving a very unmistakable indication from what customers are telling us: they value an approach to Agentic AI that spans the entire lifecycle. What customers are looking for is to put DAP to work on the extremely repetitive, mundane kinds of work that engineers handle daily—things like overseeing code, overseeing builds and deployments, and overseeing the security of their software. Take an airline, for instance—I believe I mentioned part of this during the prepared remarks—they are already leveraging DAP to take care of security vulnerability remediation automatically, dependency updates, and cloud migrations, and at this point roughly 90% of updates to component versions are being executed autonomously. Consider what that actually means. Tasks that previously needed a developer to switch context and grasp those component updates are now handled entirely by agents. Here's another case: an insurance company ran an AI hackathon using DAP and observed quantifiable gains in compliance violations, legacy modernization, the time it takes to onboard developers, and their overall security posture. Those are merely a few of the advantages customers are discovering while they assess and trial Duo Agent Platform. For a customer to make the switch and sign a committed contract as a paying customer, they need to be on a GitLab version that has support for the Duo Agent Platform. Roughly 70% of the revenue we bring in comes from supporting customers who run self-managed deployments. This isn't comparable to a cloud-native, multi-tenant SaaS offering. GitLab has an extremely varied portfolio, with 70% hosting their own infrastructure and running GitLab on their own premises, and that includes plenty who operate in air-gapped environments that are unable to connect to the cloud at all. For those customers, it generally takes around six months before as much as 50% of them are on a version such as the 18.8 release, which is the one we just announced Duo Agent Platform GA with. That's the initial timeline you should bear in mind. Naturally, the 30% of customers who run in our multi-tenant cloud are able to begin adopting right away, and we are seeing some early uptake there, but the majority of revenue won't be able to get at Duo Agent Platform for a further couple of quarters. In addition, bear in mind that committed credits really represent subscription revenue that gets recognized on a ratable basis. Even supposing we can begin turning trials into production and committed credits during the back half of the year, that money won't be recognized until later quarters, as we move into FY 28. That explains why we're hearing this early feedback from customers, and it also explains why the projections for Duo Agent Platform revenue in FY 27 are conservative.

Shrenik Kothari analyst
#26

Alright, I appreciate that thorough response. Thank you.

Jackson Ader analyst
#27

Thanks, Yao. I've got a question for Bill — over the past few quarters, with all of this enormous upheaval, has anything shifted in the way customers make decisions? What I'm really getting at is whether you've noticed any movement in sales cycles. I know I'm directing this mostly at the enterprise side, but perhaps you could just describe how customers are finding their way through this flood of change.

Bill Staples executive
#28

Yeah. When I talk to customers, they're handling this the same way most of us are — it feels like each day, each week, brings some fresh breakthrough, some new and thrilling possibility tied to AI. They're enthusiastic about what it could do, and they're working through the numerous hurdles that arrive alongside it, beginning with worries about privacy and security for their business, plus the higher costs and spending involved in putting money into AI rather than other conventional workloads. The fascinating part is that I believe a great deal of testing is underway, and clearly the initial outcomes look very encouraging in particular workloads, above all in code generation and software development. We're eager to take that initial promise of creating software through agents and transform it into genuine software innovation with Duo Agent Platform. I'm not seeing any shift in customer behavior. Our win rates are steady. As I mentioned before, our gross retention rate has just reached its highest point in 4 years. Our competitive rates are still strong. Truly, this comes down to directing our added capacity, carrying out a product strategy with more granular pricing and packaging so we can pursue that opportunity in FY 2027 with even greater strength than we showed in FY 2026.

Jackson Ader analyst
#29

Thanks.

Miller Jump analyst
#30

Hi there, wonderful. I appreciate you fielding my inquiry. Bill, near the beginning, I believe you referenced a redesign of territories. So I'm curious whether the count of accounts assigned to each rep has shifted, and what steps you take to make sure those transitions are seamless? And then regarding Jessica, did you incorporate any cautious assumptions into that guidance? Thanks.

Bill Staples executive
#31

Yes, that's a reasonable thing to ask. As you're aware, annually our territories wind up being re-divided somewhat, and that's true again this time around. My view is it will ultimately prove worthwhile; we discovered we were carrying a larger number of accounts than reps were able to handle well, and the move toward larger customers we made several years back meant the smaller segment didn't get enough attention. The new approach to territories does cut down on how many customers each rep is responsible for, and that ought to produce stronger closeness with customers, faster uptake and quicker realization of value on their side. We're putting money into technical services too, so there's more overlay assistance, notably across every segment, plus in a couple of specific areas — one being that group most sensitive to pricing, where we're adding technical services to help with their assessment, rollout, and also AI. As you know, plenty of organizations require assistance reshaping their processes for AI. Through Duo Agent Platform, there's a fresh array of technology and a fresh array of methods to apply across your software life cycle. Our technical folks are being heavily upskilled on this so customers can capture the value; additionally, we will start putting resources into forward-deployed engineers capable of stepping in to aid that adoption journey for customers.

Jessica Ross executive
#32

Yes is what I'd say in brief. As the way we take products to market keeps developing, our approach to prudence has been quite all-encompassing.

Kingsley Crane analyst
#33

Thanks for the opportunity to ask. Recently, there's been a lot of attention on modernizing legacy systems, yet if you're dealing with millions of lines of COBOL, that work doesn't happen in isolation. You'll still require version control, CI/CD pipelines, code review, and testing. Bill, given your extensive experience as a developer, how strongly do you believe we'll end up rewriting a large amount of existing code as opposed to writing brand-new code? And in what ways might this turn out to be a growth opportunity for GitLab?

Bill Staples executive
#34

Kingsley, that's an excellent question. In principle, I'd say just about any piece of software could be redone from scratch. There are tons of really exciting experiments out there where developers take code generation tools, drop them into a closed loop, and tell them to keep iterating until the resulting code emulates some other existing program. Sure, there are certain situations where that works. A discrete library, for instance, or a simple application. But take the scenario you described — say, a legacy COBOL application, possibly running on a mainframe, or some piece of enterprise software that hooks into multiple third-party SaaS backends like Salesforce or Zendesk or other systems. That, I think, is a lot harder. A lot of what gives that code its meaning lives only in the minds of the people who work with it. It was never put into writing or captured in documentation. Instead, it accumulated across a great many years, and trying to uncover those connections and hidden premises by examining the code alone is extremely difficult. When I consider where software engineering is headed, I picture something like three distinct modes that will persist for a long time to come. There are similarities to the period we all just went through with the public cloud. Think back to the beginning of the public cloud period. A major uncertainty existed: would big organizations abandon their own data centers and relocate all workloads to public cloud infrastructure? Or would they only develop fresh apps there while keeping existing systems in place? The conclusion we've reached, I believe, mirrors the direction software is heading. My expectation is that three distinct approaches will emerge. One scenario involves certain software being absolutely essential to operations — particularly for organizations in financial regulation and the public sector — where agents simply cannot be permitted anywhere near the code, whether due to security, privacy, or mere sensitivity concerns. That covers the first mode. The second mode reflects our current situation, and it's what the Duo Agent Platform enables: people and agents working together on existing brownfield codebases. Those are, you know, orchestration approaches that are coming into use right now. I think this pattern will persist for some time, especially with legacy systems like the example you mentioned. Later on, as we get better at coordinating the process and enable more iterative cycles that close the loop on code, particularly for brand-new projects built from scratch, those systems can become increasingly automated through the intelligent orchestration platform we're launching. People will end up doing less hands-on work with the code itself, other than guiding agents and managing things from a higher level. I hope that addresses your question.

Howard Ma analyst
#35

Hello, and thanks for taking my question. Bill, I'd like to hear more from you regarding the Duo Agent Platform. Through certain discussions we've had with engineering leaders, a reluctance appears to exist around bringing in an excessive number of agents from various outside software providers right away. At the same time, bigger enterprises might be weighing whether to build agents themselves or purchase them. When it comes to pricing, you now offer usage-based pricing for DAP, yet seat-based pricing remains in place for GitLab Duo Pro and Duo Enterprise, and that combination might create confusion among your customers. What's your plan for tackling these possible difficulties, and are you looking into taking the model further by merging currency-based and usage-based approaches for GitLab Duo and for DAP so that pricing and packaging feel more seamless?

Bill Staples executive
#36

Let me be clear — the capabilities of Duo Pro and Duo Enterprise already sit within Duo Agent platform. Duo Agent platform is a huge superset built on top of them. Picture it as roughly a 100x level of capability past what Duo Pro and Enterprise delivered; those packages remain on the market largely from a continuity standpoint. For any customers already planning to buy them or in the middle of doing so, we didn't want to cause disruption and make them reconsider the new platform prior to finalizing their purchase choices. Over the year to come, we'll give incentives to both customers and our field to convert those contracts into Duo Agent platform credits. Consider this a transition phase, given that Duo Agent Platform has only been in market for seven weeks. Across the next couple of quarters, it will become evident that customers view Duo Agent Platform as our AI offering moving forward.

Ryan MacWilliams analyst
#37

Thank you. I have two things to ask. Jessica, it's good to finally connect with you on this call, and I'd really like to understand how you arrived at the top-line figure in this guide, plus whether anything has shifted in the way you think about guidance. Bill, we're curious about what makes CI/CD hard for others to enter, why it represents a strategic edge for GitLab, and what would make it tough for enterprise customers to copy — say, replicating CI/CD using AI. Thanks.

Jessica Ross executive
#38

Fine. Let me begin with how I approach guidance, particularly since this marks my initial quarter as GitLab's CFO. There's a great deal going on this year. Beyond my arrival, many pieces are in motion; for GitLab, this year is about both investing and executing. As we've discussed, we're transitioning from a fee model to a hybrid one; we're inside a new product cycle with the Adapt launch, and we're simultaneously scaling how we go to market. Plenty of moving pieces. First, let me be clear: internally, our process remains the same, and it stays just as rigorous as before. During my opening weeks, I dug into the specifics of the business. That was specifically so I could back the numbers with conviction. On top of that, upon joining, I deliberately conducted a listening campaign with investors, and what came back was unmistakable. You've told us you want more clarity and openness about the way we arrive at our figures. That message came through, and I hope you can see it reflected in what we've prepared to say, and I intend to keep doing that when I give guidance. One more thing I want to say again: this business is ratable, so our board gets strong visibility, and we feed that visibility into how we operate, which means when we give you a figure, we have a lot of confidence we'll reach it. As for how the revenue guidance gets put together, I think everything that goes into it is already in my prepared remarks, but I'll say it once more. The highest net new ARR for both a year and a quarter we've ever had came in Q4 and FY 2026. Because our model can be rated, a large portion of the reduction in our guidance happens automatically; it does not represent a change in the fundamental strength of the business. The guidance captures an accurate assessment of our current position, rather than an aspirational one. Once more, let me walk through the factors behind that decline from 26%. To begin with, the mechanics. FY 2026 gained roughly 300 BPS from non-recurring tailwinds, premium price, favorable FX dynamics, and particular contract clauses that will not recur. Remove those, and the comparable growth rate brings you nearer to 23%. On top of that, because of how the ratable model works, the money we bring in right now is a reflection of booking choices that were made three years back, and over that same three-year stretch, bookings haven't grown as quickly as revenue. No matter what operational steps we take this year, that arithmetic carries straight into FY 27. Now, turning to what's reflected in the guidance: as we said before, our first order team is starting to come online, but it won't reach full ramp until the back half of the year. We've pointed out that investments of this sort take time — both before they show up in bookings and before they show up in revenue. We've built in some caution there. On top of that, we've deliberately chosen to be conservative regarding both PubSec and the mid-market — segments that were soft in Q4 and where we don't anticipate a quick recovery. Regarding GAP, as we've mentioned, customer reception has been robust, but we're still at the earliest stage of adoption. Given that this is a $1 billion business with ratable revenue, we aren't assuming any significant revenue contribution from a product that debuted only seven weeks ago.

Bill Staples executive
#39

Thanks, Jessica. I'll take a crack at the first part of what you asked, which, you know, I'd sum up as: how does GitLab's core platform hold up in a world where agents are rapidly, you know, iterating, scaling, becoming more capable and more powerful? I believe investors often get confused about this. Let me reach for an analogy, and hopefully that clarifies things. Picture the core DevOps platform GitLab provides as foundational infrastructure that humans require in order to act, and that agents require as well. Take coding agents, for instance — plenty of them can produce code, but that code has to be stored somewhere, it has to be put under version control, it has to go through testing, it has to be reviewed, and it has to be secured and validated against every standard, every compliance framework, and anything else the business is on the hook for. That infrastructure is what GitLab has spent more than a decade building. It's what companies depend on to guarantee the integrity of their software, and it isn't going anywhere. What agents bring is an artificial intelligence substitute for the human intelligence that has long been applied both to writing the code and to handling that software complexity. Think of Duo Agent Platform as that substitute for GitLab customers. They're now able to use Duo Agent Platform to automate tasks. To achieve seamless handoffs between humans and agents, and between agents and other agents, to handle every task needed to carry that software from planning all the way through deployment. Competitors can certainly provide alternative agents, but what they can't provide is a replacement for the infrastructure that is GitLab.

Moderator operator
#40

All right. Time's nearly gone. Nick, you'll close things out with the final question. Please go ahead.

Nick Altmann analyst
#41

Fantastic. Thanks. Jessica, regarding the net new ARR momentum you mentioned for Q4, it sounds like the public sector got better as the quarter went on. I realize you noted certain deals slipped during the quarter, but could you help us understand what portion of this quarter's momentum came from some of the Q3 public sector softness we observed and that sort of making it, you know, over the line in Q4? Thanks.

Jessica Ross executive
#42

No, I appreciate that question. Yeah, like you pointed out, we posted the biggest net ARR quarter we've ever had. The truth is the outcome was a blend. First order bookings came in at healthy levels, with particular momentum in Asia-Pacific. Enterprise win rates rose from one quarter to the next. Sales cycles have stayed consistent. The weakness was genuinely clustered in three places. First, PubSec, representing roughly 12% of ARR. Let me be clear on this point, our long-term thesis is unchanged. We're still the preferred partner to the U.S. government, and they keep viewing us as mission-critical. Having said that, we saw only a partial recovery after the government reopened. Certain business shifted into FY 2027, and visibility remains below the level we'd prefer. I'd say the budget situation has been noteworthy. We've seen increases across certain departments alongside considerable uncertainty in others. Once again, that's already reflected in our guide. The second area is the price-sensitive group Bill mentioned. As we told you, we estimated that at roughly 20% of ARR; it encompasses some of the SMB softness we've been talking about, along with portions of mid-market premium and customers who have less budget flexibility. Again, we believe we've sized this appropriately. We aren't observing that across the rest of the business. I'd say our 100K customers expanding 18% year-over-year and that million-dollar cohort expanding 26% have reinforced this. Enterprise ARR is still very healthy. Turning to the U.S. performance and deal slippage you mentioned, this is highly customer-specific. We had a few large deals slip from particular customers dealing with budget constraints and certain industry-specific challenges. For instance, one retailer encountered some Q4 challenges, and another large customer went through some layoffs and restructuring. These are genuine issues, and we'll keep meeting our customers where they are, but we determined this to be something quite specific. All of that has been embedded into our guide going forward.

Moderator operator
#43

Alright. Thanks. That wraps up our question-and-answer portion. We're going to be at the Morgan Stanley TMT Conference during this week, and we can't wait to see plenty of you face to face. Thanks for joining us on GitLab's 4Q and fiscal 26 earnings call. Enjoy your evening.

Read the full transcript via the API

You're viewing the first half of this call. Get the complete GitLab Inc. transcript - plus 131,000+ transcripts from 7,000+ companies, speaker segments and full-text search - through the EarningsAPI REST API or hosted MCP server.

Get an API key View API docs →

For developers and AI pipelines

Programmatic access to GitLab Inc. earnings transcripts and 131,000+ others is available through the EarningsAPI REST API and the hosted MCP server. Quarterly plans from $162 - full transcripts, speaker segments, full-text search, and the /api/v1/transcripts/recent polling endpoint for ETL pipelines.