Explorer

Atlassian's 2025 Recap: AI, Cloud, and the End of Data Center

Jan 17, 2026

If 2024 was the year Atlassian announced that AI was going to be a huge part of its future, 2025 was the year they tried to make that future feel unavoidable.

Rovo went from “interesting AI assistant” to the center of the portfolio. Jira, Confluence, Loom, Jira Service Management, Bitbucket, Focus, Talent, and the rest of the platform were increasingly described through the lens of the Teamwork Graph, AI agents, and Atlassian’s “System of Work.” And in the biggest Cloud signal yet, Atlassian finally put an end date on most Data Center products.

So yes, kind of a big year.

TL;DR Takeaways

Rovo became much more than a side feature, with Search, Chat, Agents, Studio, and Rovo Dev pushed across the Atlassian platform
Atlassian introduced and expanded several product “Collections,” including Teamwork, Strategy, Service, and Software
Data Center was officially put on the clock, with end of life set for March 28, 2029
Atlassian Government Cloud achieved FedRAMP Moderate authorization, making Cloud more viable for public-sector customers
Atlassian entered the AI browser conversation by bringing in The Browser Company, and moved deeper into engineering intelligence with DX
Jira continued its terminology cleanup, including the move from “Projects” to “Spaces”
The Atlassian Williams Racing partnership made more sense in 2025 than it did at first glance, though I still think it is very funny

Rovo Became the Center of Everything

The biggest Atlassian story of 2025 was Rovo. Not just because Atlassian talked about it a lot, though they absolutely did, but because it became the connective tissue for almost every major announcement.

At Team ‘25, Atlassian announced that Rovo would be included with Jira, Confluence, and Jira Service Management subscriptions, starting with Premium and Enterprise and then rolling into Standard. By Team ‘25 Europe, Atlassian was describing Rovo as available across paid Cloud editions and embedded throughout the portfolio.

That is a pretty big shift. In 2024, Rovo felt like a very promising, but still pretty unpolished, AI product that still needed to prove itself. In 2025, Atlassian started treating it like a core platform capability, not an add-on.

The Rovo lineup now includes:

Rovo Search: Enterprise search across Atlassian and third-party tools, powered by the Teamwork Graph
Rovo Chat: A company-aware AI chat experience that can answer questions and take action
Rovo Agents: Specialized AI teammates that can help with workflows, research, service requests, coding, project management, and more
Rovo Studio: A builder tool for creating agents, automations, assets, hubs, and apps
Rovo Dev: Developer-focused agents for planning, coding, reviewing, and automating software work

The most interesting piece to me is still the Teamwork Graph. Every company is trying to add AI to their products right now. Atlassian’s argument is that their AI is more useful because it understands the relationships between work items, docs, people, goals, services, code, incidents, videos, and third-party app data. That is a much stronger pitch than the “we put a chatbot in the sidebar” thing everyone else is doing.

As always, the question is how well it works in messy real-world environments. Permission models, stale documentation, bad Jira hygiene, duplicated Confluence pages, and weird team conventions can all make enterprise AI harder than the demo suggests. But this is the first year where I think Atlassian’s AI strategy felt less like hype-chasing and more like an actual product direction.

Collections, Collections Everywhere

Another big 2025 theme was Atlassian packaging products into “Collections.” This sounds boring, but I think it says a lot about where they are going.

The Teamwork Collection brought together Jira, Confluence, Loom, and Rovo. The pitch is that these tools form the common language for work across the whole company. Jira tracks the work, Confluence captures knowledge, Loom handles async context, and Rovo ties it together with AI.

I actually like this packaging. A lot of Atlassian customers already use Jira and Confluence together, and Loom finally has a more obvious role inside the Atlassian ecosystem. Turning a Loom recording into a Confluence page or Jira work item is exactly the kind of workflow where AI feels useful instead of gimmicky.

The Strategy Collection was aimed at leadership and enterprise planning. It combines Focus, Talent, and Jira Align. Focus is for strategic planning and tracking priorities, Talent is for workforce planning, and Jira Align continues to handle enterprise work planning. This is very much an executive/CIO product story, but it fits Atlassian’s broader goal of connecting strategy to execution.

Then, at Team ‘25 Europe, Atlassian added more to the collection strategy:

Service Collection: Jira Service Management, Customer Service Management, Assets, and Rovo for internal and external service teams
Software Collection: Rovo Dev, Bitbucket, Bitbucket Pipelines, Compass, and DX for software teams

The naming is a little corporate, sure. But the strategy makes sense. Atlassian is trying to move from “a bunch of useful tools” to “a connected operating system for work.” Whether customers experience that as simplification or just another packaging/pricing exercise will probably depend on the customer.

The potential downside is even more confusing pricing schemes. What tier of what collection do I need to get for what functionality? Atlassian’s pricing was already, IMHO, unnecessarily convoluted.

The End of Data Center Is Official

This was the one that made a lot of old-school Atlassian admins sit up straight.

Atlassian announced that most Data Center products will reach end of life on March 28, 2029. The affected products include Jira Software Data Center, Jira Service Management Data Center, Confluence Data Center, Bamboo Data Center, Crowd Data Center, Data Center mobile apps, Atlassian Data Center apps, and third-party Marketplace Data Center apps.

The key dates are:

December 16, 2025: No new Data Center app submissions
March 30, 2026: End of Data Center license sales to new customers
March 30, 2028: End of Data Center license sales to existing customers
March 28, 2029: End of life for affected Data Center products

I have very mixed feelings about this. On one hand, nobody should be surprised. Atlassian has been Cloud-first for years, Server is already gone, and all of the big AI/platform work depends on Cloud. From Atlassian’s perspective, supporting Data Center forever would slow down the platform they are trying to build.

On the other hand, I still do not like customers losing options. There are organizations with regulatory, data residency, operational, and customization reasons for staying self-managed. For them, this is not just “click migrate and enjoy Rovo.” It is years of planning, app rationalization, integration cleanup, training, governance changes, and probably a lot of budget conversations.

Atlassian is trying to soften the landing with Atlassian Ascend, FastShift, Cloud migration tooling, Government Cloud, Isolated Cloud, and a long runway. That helps. But this was still the clearest possible signal that Atlassian’s future is Cloud, full stop.

Government Cloud and Enterprise Cloud Got More Serious

The Data Center announcement would have been much harder to swallow without the other Cloud progress Atlassian made in 2025.

The biggest public-sector milestone was Atlassian Government Cloud achieving FedRAMP Moderate authorization in March for Jira, Confluence, and Jira Service Management. For U.S. federal agencies, government contractors, and other regulated organizations, that is a major door-opener.

It does not solve every regulated-industry problem. FedRAMP High and DoD IL5 are still harder milestones, and not every product is covered. But it is a real step forward, and it gives Atlassian a much stronger answer when government customers ask, “Okay, but how exactly are we supposed to move to Cloud?”

Atlassian also kept talking about more Cloud deployment options, including Atlassian Isolated Cloud, expanded data residency, customer-managed keys, higher user limits, and future compliance targets. This is the part of the Cloud story I care about the most. If they are going to force the issue with Data Center, Cloud needs to keep earning trust from the most complicated customers.

Jira Continued Its Identity Crisis, In a Good Way

Jira also kept cleaning up its language in 2025. The most noticeable change was the move from Jira “Projects” to Jira “Spaces.”

I am on record as liking this change. Jira projects were never really projects in the project-management sense. They were usually long-lived containers for a team, product, workflow, or queue. Calling them Spaces is more accurate.

That does not mean everyone loved it. Admins have years of muscle memory, internal documentation, training materials, dashboards, and screenshots built around the old terminology. And yes, there is now some ambiguity between Jira Spaces and Confluence Spaces.

Still, I think this is a net positive. Atlassian is trying to make Jira understandable outside of software teams, and terminology matters. If Jira is going to be the work system for marketing, HR, legal, finance, operations, product, and engineering, the language has to be less weird.

The Browser Company and DX

Atlassian also made two interesting acquisition moves in 2025.

The splashier one was The Browser Company, the team behind Arc and Dia. The deal was reported as a $610 million cash acquisition, and Atlassian’s stated goal is to build Dia into an AI browser for knowledge workers.

At first, this sounds random. Atlassian makes Jira and Confluence. Why are they buying a browser company?

But the more I think about it, the more it fits. A lot of work now happens in the browser. Jira is in the browser. Confluence is in the browser. Slack, Gmail, Figma, Notion, GitHub, Salesforce, and basically every other SaaS app are in the browser. If Atlassian wants to own the “system of work,” the browser is an interesting place to compete.

I am not totally sold yet. Building a beloved browser is one thing. Getting enterprises to adopt a new browser is a very different thing. But I respect the swing.

The DX acquisition makes more obvious sense. DX is an engineering intelligence company that helps teams measure developer productivity, developer experience, and AI adoption. Pair that with Rovo Dev, Jira, Bitbucket, Pipelines, and Compass, and you can see the shape of Atlassian’s software-team strategy: not just helping developers write code, but helping leaders understand where software delivery is actually improving.

The cynical take is that this could become another dashboard for executives to misuse. The optimistic take is that better engineering intelligence could help teams make smarter decisions about tooling, process, and AI investments. Both are probably true, depending on the organization.

Atlassian Williams Racing Still Makes Me Laugh

The Atlassian Williams Racing partnership became a real 2025 thing, not just a random press release. Atlassian became the team’s Official Title Partner and Official Technology Partner, with the Formula 1 team running as Atlassian Williams Racing.

I still know basically nothing about Formula 1, but I understand the branding logic more now than I did at first. F1 is an easy metaphor for high-pressure teamwork, engineering, iteration, and execution. Atlassian wants to be associated with exactly that.

Do I think Jira and Confluence are going to shave milliseconds off pit stops? No idea. Probably not directly. But as a marketing story for “teamwork at speed,” it is much less random than it felt initially.

Also, it is objectively funny to see an enterprise collaboration software company on an F1 car. I mean that affectionately.

The Business Kept Growing

Financially, Atlassian had a strong FY25. The company reported full-year revenue of about $5.2 billion, up 20% year-over-year, and Q4 revenue of about $1.38 billion, up 22% year-over-year.

Cloud continued to be the main growth engine. Q4 Cloud revenue was $928 million, up 26% year-over-year, and full-year Cloud revenue was about $3.45 billion. Atlassian also reported 2.3 million AI monthly active users at the end of FY25, then said in its Q1 FY26 shareholder letter that AI monthly active users had climbed above 3.5 million.

That is a lot of AI usage. It also explains why Atlassian is moving so aggressively. If Cloud, Rovo, and the Teamwork Graph are where the growth is, then every major decision starts to point in the same direction.

There were also leadership updates. Anu Bharadwaj announced she would depart at the end of 2025, Tamar Yehoshua was announced as Chief Product and AI Officer, and CFO Joe Binz announced plans to retire at the end of FY26. That is a lot of leadership movement during a major strategy shift.

Final Thoughts

Atlassian’s 2025 was not subtle. The company is betting hard on Cloud, AI, the Teamwork Graph, and the idea that all teams, not just software teams, should run on Atlassian.

I am still bullish, probably more than I was at the start of the year. Rovo feels like one of the more coherent enterprise AI strategies out there, and the Teamwork Graph gives Atlassian a real advantage if they can keep the data clean, permissioned, and useful.

But 2025 also made the tradeoffs clearer. Data Center customers are being pushed toward Cloud whether they are ready or not. Product packaging is getting more ambitious and more confusing. AI features are impressive, but they still have to survive contact with real enterprise messiness.

Overall though, this was probably Atlassian’s most consequential year since the Server end-of-life announcement. They did not just announce new features; they made it clear what kind of company they are trying to become.

Now we get to see whether customers follow them there.