There’s an iconic scene within the film Anchorman the place the varied San Diego information groups converge for a battle royale. Guys in ill-fitting enterprise fits and ugly ties wield chains and swords. A trident flies by means of the air. An NPR host chops off the arm of a rival with a machete. It’s nice enjoyable. Within the aftermath, over beers again on the workplace, anchorman Ron Burgundy remarks, “Boy, that escalated rapidly.”
That’s a bit how the present inventory market plunge feels. This volatility has hit tech, and particularly AI shares, more durable than most due to a fundamental monetary idea referred to as “beta.” Tech shares are sometimes excessive beta. Which means they rise greater than the inventory market in good instances and (watch for it) fall extra throughout unhealthy instances. And for those who can’t deal with the excessive beta warmth, keep out of the disruptive tech kitchen.
Whereas it’s simple to give attention to the blood and tears, these downturns are an incredible alternative to purchase shares of high quality firms, particularly fast-growing software program corporations like GitLab (GTLB) that usually commerce at a premium based mostly on future expectations. Final yr, we lastly opened a place after the corporate’s easy valuation ratio (SVR), which divides market cap by annualized income, fell to a 12-month common of 13. That was nonetheless twice the typical SVR of the shares within the Nanalyze Disruptive Tech Portfolio, however nicely beneath GitLab inventory’s wealthy valuation in years previous. The present market jitters have despatched the SVR right down to 7.7 as of right this moment (6.5 / 0.844 = 7.7 ).
Earlier than we doubtlessly pull the set off and add shares of GitLab inventory, we need to verify in with the corporate. Luckily, it simply launched its 2025 year-end outcomes, giving us the primary look into how its laborious pivot into generative AI has performed out since our earlier article. Let’s dive in.
Producing Alpha with Generative AI
As you’ll recall, GitLab’s platform unifies growth, safety, and operations (DevSecOps) right into a single utility. The platform serves as a kind of digital workspace the place builders, designers, and mission managers collaborate on software program initiatives. GitLab’s software program helps its 30 million customers observe adjustments, catch errors early, and automate repetitive duties so the ultimate product, corresponding to an app or web site, works easily. Its clients embrace everybody and every thing from solo coders dwelling of their mother’s basement to large firms and organizations like IBM and NASA (for now; extra on that later).
GitLab operates on a freemium mannequin, monetizing primarily by means of subscriptions for various kinds of enterprise packages. One large enchantment of the platform is that it’s obtainable as an everyday software-as-a–service (SaaS) providing or on-premise by means of a buyer’s personal servers or non-public cloud, very best for industries like banking or healthcare with strict information privateness guidelines. Extra just lately, GitLab intensified its give attention to generative AI by means of GitLab Duo, a set of AI-assisted instruments. This consists of GitLab Duo Self-Hosted, which allows enterprises and authorities businesses in extremely regulated sectors to deploy massive language fashions (LLMs) on non-public infrastructure or cloud companies. These AI integrations are anticipated to buoy income development and web retention price amongst enterprise clients.
Extra Seats on the Desk
And that actually appears to be the case based mostly on 2025 outcomes (GitLab operates on a kind of bizarre fiscal years, btw). Income surged 31% year-over-year (exceeding steering) to $759 million, pushed by enterprise adoption of its top-tier “Final” bundle and glossy new AI-powered instruments. As of This fall-2025, 50% of whole annual recurring income (ARR) now comes from Final, with superior safety and compliance options favored by clients in extremely regulated industries we beforehand talked about. Prospects spending greater than $100,000 per yr grew 29% to 1,229 – together with 123 purchasers representing (a minimum of) seven-figure accounts.
AI-powered Duo accounted for a couple of third of latest ARR in This fall-2025. For instance, Barclays dabbled with Duo Professional in Q2-2025 earlier than going all-in with 20,000 GitLab Duo Enterprise seats (at $39 a pop retail) in This fall, along with a corresponding variety of Final seats. Certainly, 75% of web ARR development got here from seat (i.e., person) enlargement by present clients, implying platform stickiness with a still-respectable web retention price of 123% (although down from 130% from a yr in the past). The return on funding (ROI) ought to be promoting itself.
“Prospects on our platform are having fun with 15 instances sooner time to market, 4 instances sooner function supply, and as much as 60% discount in handbook duties.” – CEO Invoice Staples
On the flip facet, new base buyer ($5,000 ARR) development was simply 10%, implying that possibly GitLab has already caught most of the large fish. Nonetheless, that doesn’t appear to be the case: Again in Q3-2023, administration mentioned GitLab served about 30% of Fortune 100 firms, suggesting there’s loads of runway so as to add enterprise purchasers – or steal them away from rivals like GitHub, the DevOps platform owned by Microsoft that had greater than 100 million builders ultimately rely. GitLab administration did spotlight a giant win involving one of many “largest cybersecurity firms in a aggressive displacement for supply code administration.”
In reality, administration mentioned its No. 1 focus is so as to add extra new paying clients in 2026 (duh!), however particularly within the mid-market and enterprise segments the place they really feel the chance for upselling and cross promoting options is strongest (duh!). Partnerships with AI cloud suppliers like AWS Bedrock and Azure OpenAI intention to seize extra of those clients, although these stellar 90% gross margins may erode a bit if AI infrastructure prices from these suppliers goes up.
An AI Engineering Buddy
On the innovation entrance, the main target is entrance and middle on generative AI, particularly as competitors on this area quickly heats up. As an illustration, Microsoft’s GitHub launched its personal AI agent by means of Copilot. Final yr, GitHub’s annual income run price hit $2 billion, with Copilot accounting for greater than 40% of the developer instrument’s income development. In the meantime, startups like Cursor AI, an AI-powered code era platform, reached $100 million in ARR final yr at warp velocity. Anthropic, a GitLab shopper, additionally markets its personal coding AI, Claude Code, making it a competitor as nicely.
Naturally, administration believes it has the technological edge. Not like standalone AI code mills like GitHub Copilot, GitLab Duo integrates AI throughout all the DevSecOps lifecycle – planning, coding, safety, and deployment. GitLAB CEO Invoice Staples notes this offers Duo an edge as a result of, “AI fashions want context. GitLab’s platform supplies it.” One wonders if Gitlab’s large code repository supplies a aggressive moat in terms of large information they’ll prepare their algos on.
Extra just lately, GitLab has launched non-public beta deployments of Duo Workflow, a next-generation AI instrument that autonomously manages code. Workflow is designed for enterprise-scale automation, positioning it as a premium upsell within the subsequent couple of years. Staples likened the AI to having a proactive “engineering buddy” that does issues like robotically repair vulnerabilities: Whereas Duo Professional flags a safety flaw, Duo Workflow robotically patches it and deploys the replace.
“Duo Workflow is our transfer from AI-assisted to AI-driven software program growth, leveraging each the ability of agentic AI and the comprehensiveness of our DevSecOps platform.” – CEO Invoice Staples
Staples makes one other fairly apparent assertion towards the tip of the presentation: “I don’t assume just a few years from now, builders will probably be constructing software program with out AI.” Most likely a greater query to ask, “Will AI be constructing software program with out builders?” That’s a query for an additional day.
What to Look ahead to GitLab in 2026
Let’s wrap it up with a few issues we’ll be watching from GitLab in 2026:
The influence of U.S. authorities cutbacks. About 12% of ARR comes from the general public sector. The corporate loved its largest quarter on this sector in Q3-2025, so Staples’ non remark, “We don’t know what we don’t know,” is a bit of troubling. Hopefully, they’ll know one thing by the following quarter.
Continued losses on the firm’s three way partnership in China, JiHu. Tailor-made particularly for the Chinese language market and its distinctive totalitarian calls for, JiHu has not precisely been profitable. The corporate misplaced $13 million on the enterprise in 2025 and expects to lose much more, as a lot as $18 million, this yr. Administration is actively attempting to “deconsolidate” the enterprise with no agency timeline in place.
GitLab once more went with conservative steering for 2026, projecting whole income of between $936 million to $942 million, representing about 24% development on the midpoint. That’s fairly near final yr’s steering (26%), so we’d nonetheless see development near 2025 if the corporate can once more exceed expectations. Money move was very robust in 2025 at practically $125 million non-GAAP working earnings, with expectations of non-GAAP working earnings of $109 million to $114 million this yr.
Nonetheless, every thing factors to a slight deceleration in development, suggesting any large windfall from the AI hype is mid-term at finest.
Conclusion
Generative AI is quickly altering the way in which many firms do enterprise. Software program growth will undoubtedly expertise among the largest seismic adjustments from automating and deploying code at scale utilizing AI brokers. GitLab seems to be positioning itself for that second.
Nonetheless, we have now to marvel about its lack of upper penetration into the world’s largest enterprises. Is Microsoft’s GitHub already too nicely entrenched given the wide-scale adoption of its suite of different software program instruments? Will startups with cheaper, extra highly effective AI options disrupt the DevSecOps mannequin in methods we are able to’t predict? Will common synthetic intelligence make all of this – together with builders themselves – moot? These are the kinds of existential questions that hold us from placing all of our cash into tech shares by allocating most of our funds for dividend development investing. In any case, you possibly can’t digitize a McD’s cheeseburger, proper? Proper?!
