We've got a great dinner tonight, too. It's super casual on the lawn. We're working on some instruments. We're working on a guitarist, but it might have to wait till next year. We're just waiting for Atlassian, so I can keep talking, or you can ask questions.
While we are waiting, what is your views on all the-
Yeah, I want to welcome James onstage, newly appointed CFO of Atlassian. Thanks again for coming. Martin's in the back. He's been with the company for many years and been a big supporter of our team. Thanks again for doing this.
Well, thank you, Brent, for having me. Hopefully, you guys can hear me okay, and thank you all for being here today.
Yeah. James spent 13 years at LinkedIn, including five years as the CFO, and he basically took the company from $10 billion - $18 billion. An incredible trajectory. Congrats on that. For those that don't know you as well, maybe talk about why you made the jump over. It's been only a few months, right? Maybe just give us your first impressions, kind of first observations, and then ultimately what you're really focused on over the next year.
Yeah. As Brent said, it's been coming up on two months at Atlassian. I would say there's no major surprises per se, but I would say I'm seeing more and more evidence in terms of what drove my conviction to the opportunity. I'll speak to some of those areas. I think the first is just how I underappreciated, I think, how well-diversified the business really is. For those maybe not following Atlassian closely, you could sort of pigeonhole it into a technology point solution for engineers, right? While there's an incredible legacy of product-led growth that really started in that domain, when I looked much more closely to the opportunity and where it's at now, and frankly, where it's been for several years now, is that Jira users, as an example, 65% of Jira users are knowledge workers. These are non-eng, non-dev.
These are folks in teams sitting across HR teams, finance teams, legal teams, and marketing teams. That number is even higher for Confluence at 70%. For JSM, our Service Collection, that's 75%. An incredibly strong and diversified base in terms of the opportunity set across enterprises and companies. That diversification, I think, also shows up in other dimensions in terms of the scale of the types of customers that we have, right? Over 350,000 customers, everyone from SMB all the way to enterprises as well. The other aspect is when I think about fundamentally where AI is now and where it's going, and the companies that are going to be incredibly well-positioned to deliver that type of value for customers durably, right? A couple of things are happening. On the AI side, it's reducing the barrier to entry as it relates to coding, right?
You're going to see a proliferation of code. We're already seeing that coming out of developers, and you're seeing a proliferation of code coming out of non-developer teams as well, marketers, legal, finance, et cetera. You're seeing more innovation, more software, and more applications being built day to day. We see that in the hundreds of millions of workflows that flow through our platform. That's doing two things. One, it's creating a lot of proliferation as it relates to the need for planning, tracking, governance, and collaboration. That's exactly where Atlassian's platform sits. The other piece is then that not all AI is created equal, right? When you think about the agents in AI that are much more intelligent, they're going to need context. I think we're seeing and hearing more and more of that right now.
When I think about Atlassian's 20-year history and the investments that we've made, not just in applications, but more and more on the Teamwork Graph that we've talked about, is that we're able to deliver higher quality outcomes and outputs through that Teamwork Graph for our customers at a much better rate as well. Those are some of the things that really stood out to me, Brent.
This concept that AI can massively disrupt your end market—these stats are obviously really good to see because Joe in customer service and Benny in accounting don't really have a choice of, you know, ripping something out.
Right
are built there. What else has been maybe surprising to you in terms of the sentiment? I think you look at those stats, and that it's pretty telling that you're more installed and deeper ingrained in these workflows than I think people understand. Are there other things that have kind of stood out to you that perhaps the casual observer right now is just getting, Hey, everything's getting AI'd?
Yeah.
What else is standing out? This concept of the work graph and all the different applications you can use, the multiple products that you're adopting, I would assume if you're adopting multiple lines of your service, that's harder to AI, if you will. Those are just a couple observations from my side.
Yeah. Ultimately, I think the proof is in the performance, right? When I look at the outperformance that we had in Q3, it came from a couple different areas. The first was cross-sell into our Teamwork Collection. That outperformed our expectations. That's really important because that's where it's the best vehicle to buy AI and unlock AI with Atlassian. When customers think about AI right now, they can certainly buy single SKUs, Confluence and Jira, and our service management tools. I think what we see over time in terms of the customer journey is that they go from single SKUs into a collection. Then they're seeing more and more value out of that collection. They buy more systems of work, right? Teamwork Collection, Jira Service Management into Service Collection. We unlock 10x credits for our customers.
It's the best vehicle for customers to actually unlock AI for their organizations right now. We're seeing that in our ARPU uplift when we monetize through Teamwork Collection and our other collections right now. That's an area where we're seeing strong performance; we're seeing really strong traction, and that continues to play through. In terms of other areas of opportunity and maybe less so surprise, we are still very early in the enterprise journey. We've been sort of evolving that go-to-market motion over the last several years, and it continues to mature. Brian and team are doing a fantastic job of that. We shared some of these stats out at our investor forum a couple of weeks ago. The million-dollar customers cohort has grown 6x over the last four years. They're growing at 39% year-over-year.
Our $3 million+ cohort has 10x'd over that same time period, growing 54% year-over-year. Our fastest-growing cohort is our $10 million+. As customers are continuing to adopt more and more AI and deploy more and more AI agents, we're continuing to take share in some of these really key markets. We talked about JSM, our Service Collection business being a $1 billion ARR business growing 30% year-over-year. Confluence is now a $1.5 billion business. Jira is a $2.5 billion business for us. A lot of opportunity to continue to grow into these different parts of the segments that we play in and all the different domains within those enterprises.
The transition to the cloud is a big catalyst; it's the end of life for the server. Maybe just talk to the dynamic. If you look at the surface, I think total revenue growth slows; the real focus should be on the cloud, right?
That's a big part of the reason why we disclosed ARR recently. To really just showcase this overall business in the right light. It removes a lot of the noise as it relates to Data Center . For those maybe not following as closely, what's happening on the data center side is that in September of 2025, we announced our data center end of life. What that did was create a couple of dynamics. One is there's ASC 606 rev rec that got pulled up from future periods into FY 2026. That's one dynamic that we saw. Q3, our most recent quarter, is our largest install base in terms of expiry base for Data Center . We had a pricing change going into effect in March of this year.
What we saw was that for customers that know that the migration to cloud is going to be a multi-year journey because of all the custom builds, all the change management, and the very large user base that they need to make sure they migrate in the right way to preserve business continuity. We saw them pull up their business, right? They expanded and said, Look, we're going to continue to commit to Atlassian. We're expanding, and we want to take advantage in the quarter. So we saw a lot of that revenue also get pulled up into the period. The other dynamic that we saw then is for customers that are actually getting ready and active to move to the cloud; they're going to see much more muted expansion, right, which makes sense. As they move to the cloud, they will continue to expand there.
Those are some of the dynamics, and that's why we wanted to show that in FY 2027. We'll see a revenue trough, a mechanical sort of revenue trough, in FY 2027 with re-acceleration out in FY 2028. The area to really point to is ARR growing north of 20% and re-accelerating as well.
When you talk about the go -to-market, Brian Duffy, I think, has done a really nice job and came from SAP, and kind of no one knows enterprise better than SAP. The limited conversations you'll let me have with him just seem like there is a lot of opportunity inside the enterprise. You mentioned, hey, it's early, but things are going well. What are the other pieces in the enterprise that you're excited about, that you're seeing that Brian and the team are doing? Maybe one thing to tie in, my question was really tied into this, is that we all do these partner checks.
Some of the partners are visibly upset, and I think it's intentional from you, where you're like, Look, we're pulling this more direct. We may hear static from the checks, but that's not really reflective of actually what is going on, and that's intentional, I believe. I think everyone wanted to hear just a view on that as well.
Yeah. It's a great question, Brent. Brian's busy closing customers, so we'll try to get him in front of you on this, Brent. I think a couple of things that we're seeing right now. If you think about the opportunity still on the enterprise side, we're in 85% of the Fortune 500, and yet it only represents 10%, roughly, of our revenue. If you think about it, then also the fact that we've got 150,000 Jira customers and roughly 65,000 JSM customers. That's 80,000 customers that we have yet to attach in terms of cross-sell opportunities across the enterprise as well. That motion is well underway right now.
There's a lot of opportunity to cross both breadth and depth as we talk about all the different cross-sell opportunities into our customers and then all the different teams across customers as well, HR, finance, legal, marketing, et cetera, as well as the developer domain. I think we're seeing a lot of opportunity there. The partner ecosystem remains incredibly important to us. It's an area where we want to make sure we continue to fine-tune and work with them to deliver the right types of value to our customer. We want to make sure that we're incentivizing the right way. Ultimately, it is really helping customers get the most out of the Atlassian platform, especially as we deploy more and more AI capabilities, making sure that both partners and our own go-to-market teams are ready to deliver that type of value and demonstrate that type of value.
From a cost perspective, obviously going through a reduction in force is painful, but you had, I think, a 10% reduction. We've seen some pretty big ones, Intuit and others. Everyone will say, Well, this is happening because the industry's starting to decay. That's the bear view. The bull view is this is going to be really profitable in a year or two years. How do you view what happened, and how do you view, from your lens, these changes we're seeing in the industry?
Yeah. It's hard to speak to other companies and their decisions that they're making right now. I think if I go back to the fundamentals of what's changing right now, right? We talked a little bit earlier about the barrier to entry for coding reducing, and that's going to continue to create more software, more innovation, and more applications. As a result, there's going to be a need for more planning, tracking, domain expertise, and governance. That's exactly where the Atlassian platform sits. So, when we saw the outperformance in Q3 as an example, that's customers telling us, Listen, we want to unlock more AI with Atlassian. That was the outperformance that we saw in Teamwork Collection. We've got now over 1,000 customers in Teamwork Collection with over 1 million seats as part of that, and that continues to grow.
The other thing that we're seeing is that it's allowing more domains to play in this space of creating more opportunity, more software, and more innovation. The other piece I'm really excited about is the Teamwork Graph, right? We showed this at our Team '26, where we put Claude Code, and we demonstrated two versions, one with Teamwork Graph and one without. Gave it the same repo, gave it the same prompt, and ultimately, we were able to show that with Teamwork Graph, we were able to demonstrate 44% higher quality outcomes at 48% fewer tokens. That's an opportunity for us to continue to deliver value to our customers, right? Not all AI is created equal. We're demonstrating that through Teamwork Graph right now.
We have the ability, because we're also an open platform, agnostic to models, to be able to optimize those models as well. There's a lot in the zeitgeist right now about token maxing and talking about making sure companies are responsibly managing those token costs, and I think Atlassian can play a big part of that.
You're a pretty global organization, and when you think about how that's playing a role, how do you think about the importance of how you've architected and built it with Australia and the rest of the world? Maybe speak to the strength of why this is working.
Yeah. I think when we think about the opportunity, it really is a global opportunity, and we've seen really healthy growth across all of the geographies that we're playing in right now from an international perspective. In terms of go-to-market opportunity, there's still a lot within the U.S., there's still a lot within APAC and EMEA. Then there are other international domains and regions that we have yet to really scratch the surface on right now. That's an area that I'm working with Brian on to make sure that we unlock over these next several years. That's one aspect from a go-to-market perspective. Then from a global domain perspective, in terms of the R&D investments that we've made, it really is at the platform level. Brent, you asked me this earlier about sort of costs and thinking about that.
We've gone through an investment cycle on the R&D side over the last several years to build an enterprise-grade cloud to make sure that's enabled and ready to also build a unified AI platform. We're starting to see a lot of leverage come out of that right now. The fact that we're able to deliver things like Rovo Dev, the MCP Gallery, and Jira Agents all within months and quarters—that's the type of acceleration and innovation that we're benefiting from through those R&D investments that we've made over the last several years. You're seeing that show up in the top line, but you're also seeing that, I think, in our cost structure as well. Our gross margins, non-GAAP, are 88%, which is three points year-over-year.
A big part of that is because we're able to continue to optimize against those platforms that we've invested in over time. We're coming off on the other end of that cycle right now. I think that answers a little bit of your question a little bit earlier, but again, we're taking a very global view of how we want to build that platform as well to be able to drive that type of leverage.
The one common thing with the partners and a lot of our investors is they look at you and they say, Okay, I got act one with Jira. I got act two with JSM. Everyone says, Okay, what's act three? Is it Rovo? Is it the whole organization taking this and the whole knowledge workers? Is it some killer new app? I don't know the answer, and maybe we don't need an answer because you have so many products today that you don't suddenly need this third superpower, but how would you describe that?
Yeah. I think if there's a third act, it really aligns with our priorities that we talked about, which are AI enterprise and system of work, and more and more driving durable, profitable growth. Let me hit on maybe a couple of those. When we think about what AI can help drive for both Atlassian and ultimately our customers, it's coming through Teamwork Graph, but we're going to meet customers where they are. Right? Customers that are very tech forward, and they've stood up internal engineering teams to build their own agents, et cetera. Atlassian is open by design. Right? Open by design, and that's how we really think about our platform. That doesn't mean we're giving away value. What it means is it's actually compounding the value.
When you think about what Teamwork Graph really does, it sort of packages all the different context that happens within enterprise workflows. Right? Right now, customers can invoke a Jira ticket, certainly through API or MCP, but they're going to get a lot more coming directly to the Teamwork Graph, right? Whether that's through their own agents or through agent to agent, like we introduced through our Jira agents. Some customers are going to want Rovo out of the box. Right? They either don't want to spend the time or resources building for that. They can get Rovo out of the box and deploy the types of agents that will serve their organizations and their teams best, whether that's sitting in marketing, finance, legal, or engineering developers as well.
There are different sorts of mediums, if you will, in terms of the types of agents and agent workflows that we can provide for customers to meet them where they are.
Your best deployment of AI internally, how would you describe that? Is it inside R&D? Is it in service? Maybe it's across the board, but how do you describe your adoption of AI internally?
Yeah. If there's maybe one thing that was a pleasant surprise for me , it's just how AI native and how AI forward Atlassians are writ large. That's not just the engineering team, it's not just the product team, it's actually across the board. Finance teams and marketing teams are going deep on Rovo, creating their own agents, and really rethinking and rewiring the workflows underneath. That's been a really pleasant surprise. I've built more agents there in my time in these last two months than I have in the last year. I think that really speaks to the culture of it. From a development perspective, where is this all going? It's showing up in our product. It's showing up in future releases and product releases like Rovo Dev, like MCP Gallery, like the Jira agents that we talked about.
Ultimately at the platform layer as well, when we think about building AI for Teamwork Graph. Teamwork Graph serves all of our collections; it serves all of our applications, and so you're seeing that come through as well.
Mike, your co-founder, has been on a lot of podcasts. I'll go on a walk and listen to it. He keeps talking about 20% growth, 20+% growth. We're in this; I just don't see the AI destruction that Wall Street keeps talking about, yet your multiple keeps fading. I guess the question I get is this conviction level and this durable growth engine, it doesn't feel like you guys are giving up on that, yet investors are questioning it right now. Maybe it's the excitement of memory and the AI party going on in infrastructure, everyone's left the software complex, which I think is partly true.
Yeah
If you keep putting up these types of growth rates with this type of margin at this multiple, I don't know how that continues. Something's got to give.
Yeah, look,
I know that's our question for the multiple.
That's right. Brent, you and the team are sort of the experts on that. What we're focused on right now is making sure that we're taking and delivering real value for our customers, and it really has to just continue to show up in the performance. I think the Q3 print really reinforced that in terms of how AI is showing up in our products, how it's showing up in our monetization as well. We're going to continue to execute, and, as I mentioned, whether it's the go-to-market enterprise opportunity and the motions that we have there or whether it's through Teamwork Graph and the acceleration in the products that we're continuing to deliver, we have a lot of opportunity still ahead of us right now. That's an area that we're going to continue to focus on in our execution.
Okay. Where the stock is now in terms of your cap allocation, how do you feel about this?
Say more, Brent?
Just in terms of capital allocation, in terms of buyback, M&A.
Yeah. Look, I think when we think about capital allocation, we first think about the opportunity that we have to serve our customers, right? Organic growth opportunities where we see the highest impact, highest ROI, and opportunity. We think about tuck-in acquisitions and opportunities there, as well, and I think we've made some of those over the past. When we think about repurchases, we've been aggressive with repurchasing in Q3. Q4 is on track right now, and really thinking about those three pillars in terms of returning capital, and making sure that we're driving durable value for our investor base, for our customers, and for our employees.
In monetizing AI, the industry is tiny today in terms of the monetization for the software industry, but when do you think this can be more measurable and more material? Is it a 2027 event, 2028?
I think different companies are on different arcs of that right now. I don't think it's one size fits all. I think what we're hearing writ large from our customers is, as it relates to AI, they do want to unlock AI. We're seeing that in our performance in things like Teamwork Collection, where they get 10x more Rovo credits, and the growth and opportunity that we see there. They're looking also for predictability, right? We're hearing a lot right now about token maxing and making sure folks are being really mindful about the tokens that they're spending right now. They're looking for predictability, and they're also looking for the ability to load balance those AI credits as well, right?
For any enterprise and organization right now, oftentimes what we see and what they see is that they'll see hyper users of those credits and then those that are deployed but not necessarily using those credits. It's helpful for them to be able to take a package of credits and tokens and to be able to load balance that across their organization as well. Right now for Atlassian, the best vehicle to monetize AI and the best way for our customers to unlock AI is through Teamwork Collection, because when they buy into that collection, they are getting 10 x the Rovo credits. We're seeing a lot of traction with that right now. We see that Teamwork Collection customers are using two times more credits than those on a standalone basis. They're deploying two times more agents, which then in turn consume more credits as well.
For Rovo customers, we're seeing credit usage grow 20% month-over-month. A lot of traction on the AI side there. It's being monetized, and we're seeing that today through ARPU in our Teamwork Collection. I think over time, that may shift over more to consumption along with seats. We also introduce our Flex program as well. This is a single wallet for customers, our largest, really strategic customers, to go all in on the Atlassian platform. That's access to the entire suite, across our collections, across all of our applications, and being able to unlock all of those credits across all of the different applications and load balance against those applications in terms of the value. Through Flex, they're able to actually drive the type of value that is most meaningful for them across the entire platform.
Thanks, James, for coming. Really appreciate your time, and it's great having you.
Awesome. Thank you, Brent.
Thanks.
Appreciate it.