The ticker symbol tells you everything. When Atlassian listed on the Nasdaq on December 10, 2015, raising $462 million at a $4.37 billion valuation — shares priced at $21, then promptly ripping 32% to close near $28 — the company traded not under some abbreviation of its corporate name but under four letters that announced a thesis about the nature of work itself: TEAM. It was, in the annals of enterprise software IPOs, an almost comically earnest choice, the kind of thing a cynical banker might have talked a founder out of. Except there were two founders, and they had been talking each other into things for thirteen years by then, and the ticker wasn't branding — it was ontology. Every product they'd built, every pricing decision they'd made, every salesperson they'd refused to hire operated on a single conviction: that the atomic unit of economic output is not the individual but the team, and that software which makes teams more effective accrues value in ways that compound beyond what traditional enterprise sales math can capture.
That conviction, expressed through a business model so unusual it baffled Wall Street for years, has since compounded into a company generating $4.4 billion in annual revenue as of fiscal year 2024, with over 300,000 customers, more than 12,000 employees distributed across 10,000 locations worldwide, and a product suite — Jira, Confluence, Jira Service Management, Trello, Loom, Bitbucket, and the AI agent Rovo — that sits at the operational nervous system of 84% of the Fortune 500. All of it built from Sydney, Australia, a city roughly 7,500 miles from Sand Hill Road, by two co-founders who started with $10,000 in credit card debt and didn't take venture capital for eight years.
By the Numbers
The Atlassian Machine
$4.4BFY2024 annual revenue
300,000+Customers worldwide
$1B+Free cash flow in FY2024 (first time)
84%Fortune 500 penetration
12,000+Employees across 10,000+ locations
$67BTotal serviceable addressable market
5,700+Apps in Atlassian Marketplace
TEAMNasdaq ticker symbol
Two Kids from Sydney Who Didn't Want to Wear Suits
The founding mythology of Atlassian is, at its core, an anti-mythology. There was no garage epiphany, no Eureka moment involving a whiteboard and a venture capitalist's checkbook. There were two twenty-two-year-olds in Sydney who graduated into the smoking crater of the dot-com bust and made a negative choice: they decided what they didn't want.
Mike Cannon-Brookes and Scott Farquhar met in 1998, both eighteen, assigned to the same scholarship program at the University of New South Wales. Cannon-Brookes — long-haired, profane, the kind of person his co-founder would later describe as "the quintessential unreasonable man" who insisted the world should work a certain way before it actually did — had the restless energy of someone who would, decades later, attempt a hostile takeover of Australia's largest carbon emitter and invest billions in a solar farm designed to beam renewable energy from the Northern Territory to Singapore. Farquhar was the analytical complement: clean-cut, precise in speech, the one who'd been offered a $50,000 job at PricewaterhouseCoopers and decided it wasn't enough — not just in money but in meaning.
Their ambitions at founding were embarrassingly modest. They wanted to earn more than Farquhar's rejected PwC salary. They wanted to not wear suits. That was it.
In 2001, Cannon-Brookes sent an email to classmates asking if anyone wanted to start a tech company. Farquhar was the only person who replied. They registered the name Atlassian — a reference to Atlas, the Titan condemned to hold up the sky, a name that operated on multiple frequencies: mythic ambition, the implication of bearing heavy loads, and the gentle absurdity of two Australian twentysomethings invoking Greek cosmology. They took out $10,000 on a credit card. They started a third-party support service for a Swedish software company, working from their bedrooms at all hours, fielding tickets across time zones.
It was not glamorous work. But it taught them something that would define the company's DNA: they hated using email to track their own developer issues. So they built an internal tool to solve the problem. Then they realized every other developer team had the same problem. The pivot from service company to product company was less a strategic decision than a recognition that the thing they'd built for themselves was the thing everyone needed.
We had two rocket engines driving us along, not just one.
The Ticket as Atomic Unit
Jira 1.0 shipped in 2002. The name was a truncation of "Gojira" — the Japanese name for Godzilla — which was itself a reference to Bugzilla, the open-source bug tracker that represented Jira's primary competition. Where Bugzilla was free but clunky, and enterprise alternatives like Microsoft's offerings were expensive and bloated, Jira occupied an empty niche: proprietary enough to be polished, cheap enough to be adopted without a procurement committee, and flexible enough to track not just bugs but tasks, projects, and — as users discovered over time — virtually any workflow that could be decomposed into discrete units of accountability.
The conceptual innovation was the ticket. A Jira ticket was an atomic unit of teamwork — a single, trackable commitment that somebody owned. It could be a bug report, a feature request, a compliance task, a pizza delivery route. The ticket didn't care. It just needed an assignee, a status, and a place to live where everyone on the team could see it. This sounds trivially obvious now, in 2025, when every knowledge worker on earth has internalized the Kanban board. But in 2002, most software teams coordinated through email chains, spreadsheets, and shouting across cubicle walls. The idea that every piece of work should be a visible, stateful object in a shared system — that was genuinely new for the price point at which Jira offered it.
Within five years, Jira had become, in the words of Accel partner Rich Wong, "the standard that you had to integrate your software with." Wong's portfolio companies — Accel-backed startups across Silicon Valley — had adopted Jira organically, without anyone at Atlassian placing a single sales call. This was the first evidence of a pattern that would repeat across every product in the portfolio: developers adopted the tool, told other developers, and the software spread through organizations like a benign contagion.
Confluence followed in 2004 — a wiki-based collaboration platform that gave teams a shared space for documentation, specifications, meeting notes, and the messy work-in-progress that usually lived in someone's head or, worse, in someone's inbox. The integration with Jira was seamless and deliberate. A Confluence page could reference Jira tickets; a Jira ticket could link to Confluence documentation. The two products created a closed loop: plan the work in Confluence, track the work in Jira, document the results back in Confluence.
"We had two rocket engines driving us along, not just one," Cannon-Brookes said. And the engines fed each other. A team that adopted Jira had a natural reason to adopt Confluence. A team that used both was structurally embedded in the Atlassian ecosystem — not because of contractual lock-in or aggressive bundling, but because the products were genuinely better together than apart.
The No-Sales Heresy
The most radical decision in Atlassian's history was the one that didn't happen. They didn't hire a sales force.
This requires context. In the early 2000s, enterprise software was a sales-driven industry. Companies like Oracle, SAP, and IBM employed armies of account executives, solution engineers, and partner channel managers who flew to client sites, ran multi-month procurement cycles, and closed deals measured in millions of dollars. The conventional wisdom — backed by decades of precedent and billions of dollars in commission checks — was that enterprise software required enterprise sales. You couldn't sell complex tools to large organizations without humans in the loop explaining the value proposition, navigating the buying committee, and negotiating the contract.
Atlassian rejected this model entirely. Not as a temporary bootstrapping compromise but as a philosophical position. Their products would be priced low enough — Jira initially sold for as little as $800 for a small-team license — that individual developers or team leads could expense them without executive approval. The products would be distributed through the company's website, downloadable and deployable without a sales conversation. The pricing would be transparent, published on the website, no "contact us for a quote" theater.
The economics were counterintuitive. Typical enterprise software companies in the mid-2000s spent 40–60% of revenue on sales and marketing. Atlassian spent roughly 15–20%. This wasn't frugality — it was a bet that the savings could be reinvested into R&D, that better products would generate more word-of-mouth adoption, and that word-of-mouth adoption at low price points would create a distribution channel more powerful and defensible than any sales team.
By the time Atlassian hit $100 million in revenue in 2012 — a milestone reached without a single traditional enterprise sales rep — the heresy had become a playbook. Silicon Valley would later call it "product-led growth" and build an entire category of venture-backed companies around the concept. But Atlassian didn't have a name for it. They just had a conviction that their products should sell themselves, and a pricing structure designed to make that possible.
Atlassian's sales & marketing spend vs. enterprise software peers
| Company | S&M as % of Revenue (circa 2015) | Primary Distribution |
|---|
| Atlassian | ~15-20% | Self-serve, product-led |
| Salesforce | ~50% | Enterprise sales force |
| ServiceNow | ~45% | Enterprise sales force |
| Oracle | ~40% | Enterprise sales force + channel |
The Eight-Year Bootstrap
For eight years — from 2002 to 2010 — Atlassian operated without venture capital. This was not entirely by design. Cannon-Brookes and Farquhar were in Sydney, 7,500 miles from Sand Hill Road, building software for developers at a time when Australian venture capital barely existed for enterprise technology. But the constraint became a virtue. Without external capital, they couldn't afford to build a sales team even if they'd wanted one. Without a sales team, they had to make products good enough to sell themselves. Without the pressure to hit venture-scale growth metrics on someone else's timeline, they could compound steadily, reinvesting profits into product development.
By 2005, three years after launching Jira, they had 1,000 customers. By 2010, when they finally took investment — $60 million from Accel Partners at an $8 billion valuation, according to various reports — they were already profitable. The Accel money wasn't needed for operations. It was, in some sense, validation capital — a signal to the market that this peculiar Australian company with no sales team was, in fact, a real enterprise software business.
The bootstrapping period shaped Atlassian's culture in ways that no amount of subsequent venture capital could dilute. It created an institutional allergy to waste, a default assumption that money should be spent on building products rather than acquiring customers, and a deep respect for the economics of compounding small amounts of revenue from large numbers of customers rather than extracting large amounts from a small number.
It also created something harder to quantify: a co-founder dynamic forged under financial pressure rather than abundance. Cannon-Brookes and Farquhar had been best men at each other's weddings, became parents three months apart, owned property next to each other in Sydney. Their partnership wasn't a Silicon Valley co-founder marriage of convenience — it was the closest thing the Australian tech industry had to a Lennon-McCartney collaboration, except both of them were still speaking to each other two decades in.
The Dual-CEO Experiment
Most companies with two co-founders resolve the leadership question by making one the CEO and the other something else — CTO, President, Chief Product Officer. Atlassian refused to choose. From founding through 2022, Cannon-Brookes and Farquhar served as co-CEOs, sharing the title, the responsibilities, and the Atlassian-Ukraine blog post that one would draft while the other handled something else.
The arrangement worked because their differences were complementary rather than competitive. Early investor Rich Wong called Farquhar "more analytical," the one who thought in systems and structures. Cannon-Brookes was the visionary provocateur, the one who said the world should work a certain way and then set about bending it. "Mike is kind of the quintessential unreasonable man," Farquhar told CNBC. "'The world should work this way.' 'Mike, it doesn't yet.'"
The practical mechanics were straightforward: they divided domains rather than duplicating them. On any given decision, one would take the lead and the other would provide honest feedback sharpened by twenty years of shared context. The overhead of a dual-CEO structure — the coordination costs, the potential for disagreement, the confusion it creates for direct reports who don't know which boss to listen to — was more than offset by the bandwidth it created. Two CEOs meant Atlassian could operate on two fronts simultaneously: product vision and operational execution, or customer strategy and cultural infrastructure, without either dimension being starved for attention.
In August 2022, Farquhar stepped back from the co-CEO role, leaving Cannon-Brookes as sole CEO. The transition was framed as natural evolution rather than rupture — Farquhar had been signaling his interest in other pursuits — but it marked the end of one of the longest-running co-CEO arrangements in the technology industry. The company that defined itself by teamwork had, for twenty years, practiced it at the very top.
The Marketplace as Moat
In 2012, Atlassian launched the Atlassian Marketplace — a platform where third-party developers could build and sell extensions, integrations, and add-ons for Jira, Confluence, and the rest of the product suite. The Marketplace solved a problem that every horizontal software platform eventually confronts: how to serve the infinite diversity of customer workflows without building infinite features yourself.
A hospital system using Jira needed different custom fields, automations, and reporting than a game studio using Jira. Rather than trying to build every possible configuration into the core product — the path to the feature bloat that would eventually become one of Jira's most persistent criticisms — Atlassian outsourced the customization layer to an ecosystem of independent developers. The Marketplace gave these developers a distribution channel (access to Atlassian's customer base), a revenue model (Atlassian took a percentage of sales), and a reason to keep building (the larger the Atlassian customer base grew, the larger the addressable market for Marketplace apps).
By 2025, the Marketplace hosted over 5,700 apps. This ecosystem created a compounding moat: every app built for the Atlassian platform increased the platform's value to customers, which attracted more customers, which attracted more developers, which attracted more apps. The flywheel was a textbook example of platform economics — Atlassian's version of the iOS App Store, except for enterprise workflows rather than mobile games.
But the Marketplace also created a tension that would grow more acute over time. Third-party developers built businesses on the Atlassian platform, invested resources in maintaining their apps, and accumulated customers — all of which made them dependent on Atlassian's platform decisions. When Atlassian made the strategic choice to migrate its entire customer base from on-premises Server deployments to Cloud, the Marketplace ecosystem had to follow, at significant development cost, or die. Some didn't survive the transition.
Acquiring the Adjacent
Atlassian's acquisition strategy followed a consistent logic: buy products that teams already use alongside Atlassian's own tools, integrate them into the platform, and let the combined value accelerate the flywheel. This was not empire-building for its own sake. It was a systematic effort to own more of the team's daily workflow.
🔗
The Acquisition Timeline
Key deals that expanded the Atlassian platform
2007Acquires Cenqua — gains Bamboo (CI/CD), FishEye, and Crucible (code review), deepening the developer toolchain.
2010Acquires Bitbucket, a Git-based code hosting platform — Atlassian's answer to GitHub, offered free private repositories that attracted cost-conscious developers.
2012Acquires HipChat, a team messaging tool — Atlassian's first move into real-time communication.
2017Acquires Trello for $425 million — gains a massively popular Kanban-style project management tool with tens of millions of users, most of them non-developers.
2018Acquires OpsGenie, an incident management platform — extends into IT operations and service management.
2023Acquires Loom for approximately $975 million — adds asynchronous video messaging to the collaboration stack.
The Trello acquisition was the most strategically significant. Trello's user base was enormous but largely non-technical — project managers, marketing teams, event planners, freelancers. It represented Atlassian's clearest bet that its future lay beyond the developer persona that had built the company. Trello users were the wedge into work management, a market that Atlassian would eventually define as one of its three core strategic pillars alongside software development and service management.
The HipChat story, by contrast, was a cautionary tale. Atlassian acquired HipChat in 2012, launched its successor Stride in 2017, and then — less than a year later — conceded defeat to
Slack, selling both HipChat and Stride's intellectual property to Slack in exchange for an equity stake. The retreat was uncommon in enterprise software, where companies typically cling to underperforming products for years rather than admit a competitive loss. Cannon-Brookes and Farquhar's willingness to exit a market they couldn't win was as revealing as any of their successes: the discipline to stop throwing resources at a losing position and reallocate to higher-return opportunities.
The Server Sunset and the Cloud Bet
The most consequential — and most controversial — strategic decision in Atlassian's recent history was the forced migration from Server to Cloud.
For most of its existence, Atlassian sold its products as downloadable software that customers installed and ran on their own servers. This "Server" deployment model was the foundation of the company's original product-led growth engine: teams could download Jira or Confluence, install it behind their firewall, and start using it immediately. The low price point and self-service deployment created the friction-free adoption that defined Atlassian's distribution advantage.
But the Server model had structural limitations. Customers on self-hosted deployments were running different versions of the software, making it expensive for Atlassian to maintain backward compatibility. Upgrades were the customer's responsibility, which meant many ran years-old versions with known security vulnerabilities. Atlassian had limited telemetry on how Server customers actually used the products — data that was essential for improving the software and developing AI capabilities. And the economics were less favorable: a one-time license fee generated revenue once, while a Cloud subscription generated revenue every month.
In 2020, Atlassian announced it would end support for Server products, effective February 2024. Customers had to migrate either to Atlassian Cloud — hosted and managed by Atlassian — or to Data Center, the company's self-managed enterprise offering, which was priced significantly higher than Server had been.
The decision was strategically brilliant and operationally painful. It consolidated Atlassian's customer base onto a platform the company controlled, enabling faster feature releases, AI integration, and the kind of usage data that powers modern product development. It transformed the revenue model from one-time licenses to recurring subscriptions. And it created a clear path to the enterprise market, where CIOs increasingly preferred cloud-hosted software to the maintenance burden of on-premises deployments.
But it also forced hundreds of thousands of teams through a migration they didn't want, at a cost many hadn't budgeted for, on a timeline they didn't choose. Some customers left entirely. The Marketplace ecosystem was disrupted. And the trust that Atlassian had built through decades of low prices and customer-friendly policies took a hit — "Don't #@!% the customer," one of the company's five core values, became a punchline in some corners of the developer community.
We successfully wound down support for Server; introduced Rovo, taking human and AI collaboration to the next level; and ushered in the next era of a unified Jira.
The Feature Bloat Paradox
Success breeds features, and features breed complexity. This is the paradox that every successful horizontal software platform eventually confronts, and Jira has confronted it more visibly than most.
The same flexibility that made Jira universally adoptable — its ability to track any workflow, in any domain, with any level of customization — gradually turned it into what former Atlassian engineers described as a "Swiss Army knife that nobody knows how to use anymore." A product designed to give teams clarity became, for many users, a source of cognitive overload. Custom fields proliferated. Workflow configurations grew labyrinthine. The gap between what Jira could do and what any individual team needed it to do widened with every release.
"You're not building for your existing customers anymore," one former Atlassian engineer observed. "You're building for the imagined customer that might come in the future. And that's when you lose focus."
The tension was structural. Jira's competitive advantage was its flexibility — the ability to serve a three-person startup and a 50,000-person bank from the same platform. But flexibility, taken to its logical extreme, becomes complexity. And complexity is the enemy of the very thing Atlassian claimed to sell: team effectiveness. The company's own research found that 65% of knowledge workers said responding to notifications was more important than making progress on top priorities — a finding that indicted not just email and Slack but, implicitly, the notification-heavy, ticket-saturated workflows that Jira itself enabled.
Atlassian's response was to simplify aggressively — unifying Jira Software and Jira Work Management into a single Jira product in 2024, stripping away redundant interfaces, and investing in AI-powered features (through Atlassian
Intelligence and the Rovo agent) designed to surface relevant information rather than requiring users to navigate to it. Whether the simplification can outpace the complexity is an open question. Product entropy is a law of software physics, and Jira has been accumulating features for twenty-two years.
Team Anywhere, or the Office as Optional
In 2020, before the pandemic made remote work a universal experiment, Atlassian — whose products were literally the infrastructure of remote collaboration — committed to what it called "Team Anywhere." The policy allowed all 12,000-plus employees to work from wherever Atlassian had a legal entity, as long as their time zone was compatible with their role and their team leader approved.
This was not pandemic improvisation. "We were always experimenting in this space heavily," said Avani Prabhakar, Atlassian's chief people officer, "because it sits very well with what we do for our customers." The company built a dedicated research function — the Teamwork Lab, staffed with behavioral scientists — to study distributed work not as an HR accommodation but as a product problem. How do you maintain knowledge transfer without hallway conversations? How do you create serendipitous collaboration when people aren't in the same building? How do you prevent the information silos that form when work happens in private channels rather than shared spaces?
Their answers became product features. Confluence's "open by default" philosophy — where all documents are visible to the entire company unless specifically restricted — was both a cultural practice and a distribution mechanism. New employees discovered that their rough drafts were visible to the entire organization, felt the initial vulnerability of working in public, and then experienced the payoff: comments from strangers across the company who'd worked on similar problems, shortcuts that saved weeks of futile effort, connections that would never have happened in a hallway.
For us, it was never a thing where we will think about return to office one day when things get normal.
The results, as Atlassian reported them, were striking: 92% of employees said Team Anywhere helped them do their best work, 91% cited flexibility as a primary reason for staying, and the company's workforce tripled during the distributed work era while candidate acceptance rates rose 20%. The number of candidates applying for open roles more than doubled.
The skeptic's response is that Atlassian was dogfooding its own products — of course a collaboration software company would claim that distributed collaboration works. But the data was harder to dismiss than the narrative. Atlassian's revenue grew from roughly $1.6 billion in FY2020 to $4.4 billion in FY2024 during the Team Anywhere era, and free cash flow crossed $1 billion for the first time. Whatever the distributed model cost in hallway serendipity, it didn't show up in the income statement.
The Enterprise Ascent
For most of its history, Atlassian sold to teams, not to enterprises. A developer team would adopt Jira, a marketing team would adopt Trello, an IT team would adopt Jira Service Management — and the company would grow one team at a time, bottoms-up, with no enterprise sales rep orchestrating the expansion. This was the product-led growth model that had taken Atlassian from zero to $4 billion.
But there was a ceiling. Or rather, there was a gap: 84% of the Fortune 500 were Atlassian customers, yet those enterprise accounts represented only 10% of total revenue. The average Fortune 500 company was spending a fraction of what it could on Atlassian products. The potential within the existing enterprise customer base alone, by Atlassian's own estimate, was $14 billion.
Closing this gap required capabilities that Atlassian had historically disdained. Enterprise customers needed compliance certifications (FedRAMP, SOC 2), data residency controls (the ability to keep data in specific geographic regions), advanced administration tools, single sign-on integration, and — most heretically — human sales interactions. You don't sell a seven-figure platform deal to a Global 2000 CIO through a website checkout page.
Atlassian's approach was characteristically idiosyncratic: build enterprise-grade capabilities into the platform, achieve FedRAMP "In Process" status (announced in FY2024), add data residency in six additional regions, and layer a modest enterprise sales motion on top of the product-led foundation rather than replacing it. The company was not abandoning its no-sales DNA — it was adding a sales layer for the subset of customers whose purchasing process demanded it, while keeping the self-serve engine running for everyone else.
The tension was real. Every dollar spent on enterprise sales was a dollar not spent on R&D. Every enterprise feature added complexity to a platform that was already struggling with feature bloat. And the cultural implications were significant: a company that had defined itself by the absence of salespeople was hiring them. The question was whether Atlassian could add an enterprise motion without losing the product-led soul that had made the company possible in the first place.
The AI Wager
In April 2023, Atlassian introduced Atlassian Intelligence, built on its own machine learning models plus OpenAI's technology. The feature embedded AI capabilities across the product suite — summarizing Confluence pages, generating Jira ticket descriptions, powering virtual agents in Jira Service Management. In 2024, the company announced Rovo, described as an "AI teammate" capable of searching across the entire Atlassian platform and connected third-party tools to surface relevant information, answer questions, and automate routine tasks.
The AI investment was both offensive and defensive. Offensively, AI represented the possibility of making Atlassian's products dramatically more useful — reducing the cognitive overhead that had accumulated over two decades of feature accumulation, transforming Jira from a tool you had to navigate into a tool that navigated for you. If Rovo could surface the right Jira ticket, the right Confluence page, the right Loom video at the right moment, the feature bloat problem would be partially solved by an intelligence layer that abstracted away the complexity.
Defensively, AI posed an existential question: if AI agents could eventually manage projects, write documentation, and triage service requests autonomously, would teams still need Atlassian's products? The company that built software for teamwork was betting that AI would augment teams rather than replace them. The data inside Atlassian's platform — years of tickets, pages, incident reports, and decision records — was the training data for AI that could understand how specific organizations actually work. No competitor, not even Microsoft with its vast resources, had access to that particular corpus.
The cloud migration suddenly looked prescient. Server customers, isolated behind their firewalls, would have been unreachable by Atlassian's AI capabilities. Cloud customers, all running the same platform version, all generating data that Atlassian could (with appropriate permissions) use to train and improve AI models, were the foundation on which the AI bet depended. The Server sunset hadn't just been a revenue model transition. It had been a prerequisite for the AI era.
The Values as Operating System
Most companies list their values on a careers page and promptly ignore them. Atlassian's five values — codified in 2007, when the company was still small enough that values could be enforced by proximity — functioned as something closer to an operating system, shaping decisions in ways that were visible in the product, the pricing, and the organizational structure.
Open company, no bullshit. This was Confluence's "open by default" philosophy made cultural. All documents visible. All work-in-progress exposed. The vulnerability of working in public — an executive drafting a strategy document that any engineer could comment on — was not a bug but a feature.
Build with heart and balance. This manifested in the company's distributed work policy, its investment in employee well-being research, and its refusal to adopt the growth-at-all-costs posture that defined many Silicon Valley peers.
Don't #@!% the customer. The most violated of the five, at least in the eyes of customers who were forced through the Server-to-Cloud migration. But also the most consequential: Atlassian's low pricing, transparent pricing pages, and resistance to the enterprise software industry's "call us for a quote" theater all traced back to this principle.
Play, as a team. The ticker symbol. The co-CEO structure. The product philosophy. The conviction that the unit of output is the team, not the individual.
Be the change you seek. Atlassian's Pledge 1% commitment — donating 1% of profit, employee time, and equity — was cofounded by the company in 2006. Cannon-Brookes' personal climate activism, including his attempts to transform AGL Energy and his investment in the Sun Cable solar project, extended this value beyond the corporate context into something more personally radical than most tech billionaires attempt.
The values were not decorative. They were load-bearing. They explained why Atlassian didn't hire salespeople, why it priced transparently, why it bet on distributed work before a pandemic made it fashionable, and why its co-founders split leadership rather than competing for it. Whether they would survive the pressures of the enterprise transition — the compliance requirements, the sales motions, the complexity of serving governments and regulated industries — was a question the next chapter of the company would have to answer.
$67 Billion of White Space
Atlassian's FY2024 shareholder letter framed the company's opportunity in three markets: software development, service management, and work management. Combined, these represented a $67 billion total serviceable addressable market growing at 13% annually. The company's $4.4 billion in revenue suggested roughly 6.5% market share — significant enough to be a major player, small enough to support the growth thesis for years.
The three markets were not independent. Software development teams used Jira and Bitbucket. IT service management teams used Jira Service Management and OpsGenie. Work management teams used Jira, Confluence, Trello, and Loom. The lines between these markets were blurring — a trend Atlassian called the convergence of "dev and IT" — and the company's cross-platform integration was designed to be the bridge. An incident in Jira Service Management could trigger a Jira ticket for the development team, link to a Confluence page documenting the root cause, and generate a Loom video explaining the fix. The platform's value increased with every product a team adopted, and every product a team adopted increased the switching cost.
The first $1 billion revenue quarter arrived in Q4 of FY2024.
Free cash flow exceeded $1 billion for the full year. The company had 300,000 customers and was adding more. The growth was not hypergrowth — Atlassian had never been a "grow at all costs" company — but it was steady, profitable, and compounding in the way that only a platform with genuine network effects and high switching costs can compound.
And yet the most telling number in the shareholder letter was not the revenue figure or the cash flow milestone. It was this: within Atlassian's existing enterprise customer base, the company had identified $14 billion of revenue potential. The land was already grabbed. The expand — selling more products to the teams already inside those enterprises, moving from departmental tool to organizational platform — was the next decade's opportunity.
Within our existing enterprise customer base alone, we have identified $14 billion of revenue potential. Today, 84% of the Fortune 500 are Atlassian customers, yet they represent only 10% of our total business.
On the day Atlassian's stock debuted in December 2015, at $21 per share under the ticker TEAM, the company was worth $4.37 billion and had never generated $1 billion in annual revenue. By fiscal year 2024, it was generating $1 billion in a single quarter. The two guys from Sydney who started with credit card debt and a refusal to wear suits had built a company whose products now operated on two planets — Jira, it turned out, tracked issues on the Mars Perseverance rover. Somewhere in the Atlassian Platform, there is a ticket for driving a robot across the surface of another world, and someone is assigned to close it.
