5 min read

Razor's Edge: Long Gtlb

I've always skewed skeptic on this one, but the GenAi coding boom got me interested in it. Initially, you hear cursor/windsurf metrics and you think GTLB terminal value is permanently impaired and it's a short. Then you start doing some work and you realize maybe the opposite is true.

Summary Bullets of My Thinking on This

  • Definitely viewed as a platform for devsecops vs. point product
  • Devsecops is structural bull mkt
  • Presently overshadowed by genAi coding ide's like Cursor and Windsurf and seat driven model concerns
  • 25%+ grower (printing 20% FCF margin) still in ISV land trading at 6x EV/Sales makes it an extreme valuation outlier

I think one can have a good discussion about where things are at in GenAi in certain spaces and kind of go back to TAM convos of the past.

Everyone knows subscription helped Adobe's TAM expand though social I think also huge deal there too, but with GenAi Adobe draws all kinds of criticism because the perception is the end market for creatives shrinks and also competitive landscape has exploded.

Now we know that in a sense the creative end mkt argument is not true as just about anyone on earth is now capable of some impressive creative copy, so really the point here is the paid market and competition has changed enough to impair the core value prop of the incumbent leader in creative software.

Coding assistants don't feel the same way. They feel more like the invention of pocket calculators to me right now. And you could argue they open up coding to a lot more people which maybe has positive temp cycle affect on the human devops tam vs shrinking it. You can also argue that the IDE's enhance platform value. I starting thinking this as I was doing work over the past week. There seems to be a lot more end customer takes that genai code assistants' cursor/windsurf are tam enhancers for github/gitlab w a skew towards gitlab over privacy/interops advantages.

Then mgmt did get asked to comment on this on the call and had this to say which naturally had some serious confirm bias for me:

I know it's also sometimes confusing about where 1 tool begins and another tool ends, but the way I think about this is cursor and wind search since you mentioned those 2, they're helping engineers create applications and new code, in particular.

And GitLab actually picks up where those code assistants leave off. All of the code that they generate does need all of the things that GitLab does so well, testing, securing, analyzing, packaging, building, deploying, right? And so they already actually work great with GitLab. Cursor works great, Windsor works great, and they already are driving code into GitLab.

And so that's why a lot of our customers do test those tools in addition to duo. And we do offer co-creation capabilities as well. As I mentioned, the Duo code suggestions capability that's now part of every premium and ultimate license has the same ability as a cursor or windsor in multiple IDEs development tools to create code and create applications, and our capability there is only going to get stronger over time.

But we really believe in interoperability and being an open platform. And so we embrace those tools. We're excited by the innovation of brand developers and we don't see them challenging the domain, the parts of the DevSecOps life cycle that we are world-class at today. So I guess I'll just finalize this by saying, we also -- as we think about our AI strategy, we're just thinking much broader from the start.

It's a multipronged strategy that looks across the software life cycle, and we're really excited about the approach we've brought to the market already with agentic AI with Amazon Q and our partnership there as well as the Duo workflow product and the beta that's now playing out with GA to come. So as I think about the broader picture of what we're doing versus what the startups are doing, I think they're really complementary and really exciting for customers down the road. I hope that answers your question.

So, ok DUO is not really tool of preference for GenAi IDE, but that doesn't really matter big picture and it will get better. Also, seems those customers testing duo workflow (agentic AI for devsecops) are way more excited about that which will be GA by year end. That to me is a intriguing setup narrative wise. ....

Terminally impaired trading valuation priced in when evidence exact opposite might be true.

EARNING/REACTION

I got asked this morning about why I seemingly have started bottom fishing or maybe going into value-trap land more in SaaS vs owning the cleanest names.

It's a fair question as I do think you should avoid complicating things.

But just looking at my saas p/l on the long side last 12 months the top contributors are RBRK/SNOW/NET/HUBS/MNDY/OKTA/CRWD/TEAM/NOW/NTNX. Yea, a lot of it was very nice timed event trading in these, but outside of TEAM/HUBS the rest make up the chosen one club X-PLTR. My biggest drag had been CRM, but i traded the print well and now its basically WDAY. Short side ESTC and ORCL been the stars with no drags thankfully.

Guess the point is you eventually have to look for more unique things as the easy stuff all became consensus, and the track record been pretty good. That said I actually think this Gitlab has a strong case on its side here as being in the easy camp.

The argument for why it has dropped on earnings..

  • narrowest beat since its been public
  • lowest base customer adds as public company
  • NRR sequential decline of 1%
  • no annaul guide raise

Ok, this is game within the game stuff that I at times lose my patience with but need to remind myself this is part of the process.

Why would I lose patience?

Well, this thing is down to 6x ev/sales after printing 27% growth Q and i think i read autopsies on the buy-side bogeys from 5 people immediately several of which never even used to speak this language. 6X sales with this growth and margins is basically a distressed asset in public SaaS these days. So, you go back to part one and have the terminal convo because all this other stuff is a waste of time at this price. But let's do it anyway...

If I wanted to counter some of the bogey decel stuff I'd point out

  • this was a very tough seasonal comp for them based on a exceptionally strong Q4
  • sub was actually strong vs license which is supposed to be a good thing
  • 100k adds were solid
  • pricing changes which are a focus a offer positive upside ahead for these companies as they layer in consumption did impact lower end
  • nrr calendar adjusted was flat
  • new mgmt team likely conservative which is boiler plate at this point
  • Federal which big deal here and has headwinds was. good for them which says a lot on platform value

Anyway, buying this down here has worked like clockwork and I now think you actually have more strategic upside in this name. Suitors now have to be half a dozen players vs usual google rumor which means even the most non-commital of traders has to love the risk/reward on buying this dip. Also, you have a product event coming up in 2 weeks which is always good for one of these.

My general thinking is you have a low hanging fruit 20% here with 50% to 100% upside if the narrative turns. That's a nice rent-to-own option.