Atlassian Corporation
TEAM 6.5Price and target zones价格与目标区间
What the company does公司做什么
Atlassian makes the tools software teams use to plan and track their work. Its main products are Jira, for tracking tasks and projects, and Confluence, for team documents. Most software teams use at least one.Atlassian 制造软件团队用来规划和跟踪工作的工具。它的主打产品是 Jira(用于跟踪任务和项目)和 Confluence(用于团队文档)。大多数软件团队至少会用到其中一个。
Earns from营收来源 subscriptions. In FY2025, 95% of the $5.2 billion revenue was subscription.订阅。2025财年,52亿美元收入中有95%来自订阅。
Read the full description查看完整介绍
Atlassian Corporation provides a collaboration software that enables organizations to connect all teams through a system of work that unlocks productivity at scale worldwide. The company's product portfolio includes Jira, Confluence, Loom, Jira Service Management, Rovo, Bitbucket, Compass, Jira Product Discovery, Focus, Talent, Trello, and Guard. It also offers Collections, a curated sets of apps and agents built on the Atlassian cloud platform and designed to solve cross-functional customer workflows, including teamwork, service, strategy, software, and product collections. The company was founded in 2002 and is headquartered in Sydney, Australia.
Revenue breakdown营收构成 share of revenue营收占比
Atlassian's core business. Customers pay a recurring fee to use Jira, Confluence, and related tools in the cloud, or through term licences for on-site Data Center deployments. This income is steady and easy to forecast, and it lets Atlassian invest steadily in new features.Atlassian的核心业务。客户按经常性费用使用Jira、Confluence及相关工具的云端版本,或通过现场部署的Data Center产品的期限许可协议使用。这部分收入稳定且易于预测,使Atlassian能够持续投资新功能。
- Jira — Plans and tracks project and software work.规划和跟踪项目及软件工作。
- Confluence — A shared workspace for team documents and knowledge.供团队存放文档和知识的共享工作区。
- Jira Service Management — An IT and service help desk.IT 和服务台。
Revenue outside the main subscriptions. Third-party developers sell add-on apps for Jira and Confluence through the Atlassian Marketplace, and Atlassian takes a cut of each sale. This category also includes older on-site software licences and advisory and training services. It builds a wider ecosystem around Atlassian's core tools, which raises the cost for customers to switch away.主要订阅业务之外的收入。第三方开发者通过Atlassian Marketplace销售Jira和Confluence的附加应用,Atlassian从每笔销售中抽成。此类别还包括较旧的现场软件许可以及咨询和培训服务。它围绕Atlassian的核心工具构建了更广泛的生态系统,从而提高了客户的转换成本。
Company facts公司资料
- Founded成立时间
- 2002
- Headquarters总部
- Sydney, Australia
- Employees员工人数
- 13,301
- Industry行业
- Software - Application
- Country国家
- Australia
Investment thesis投资逻辑 from the watchlist来自关注列表
🟡 RE-ANALYZED 8/07 — HOLD. Score unchanged, every input behind it changed. Zones fully re-derived (the old Add <$68 sat 123% below spot → [[pitfall-stale-entry-zone-suppresses-a-name]]). The April thesis was right and has been PAID: +114% since the 4/26 analysis at $71.55, +173% off the $56.01 low, and +38.9% on 8/07 alone — the largest one-day move in company history.
What's working
- Q4 FY26 was the best print in the company's history: rev $1,766M +28%, Cloud +31% and ACCELERATING (26→26→29→31 straight through the SaaSpocalypse), RPO +44% to $4,817M vs cRPO +27% (customers signing longer and larger exactly as the AI question peaks — the hardest datapoint to explain away), non-GAAP EPS $1.87 vs $1.50, GAAP op income positive for the first time since FY22 (Q4 +$210.7M, 12% margin; first profitable quarter ever). $5M+ ARR customers +70% to 69; $3M+ +50% to 164. 🔍 THE HEADLINE "MARGIN CUT 30%→25%" IS FALSE and the letter says so on p.20 (verified verbatim against the PDF): FY26 was flattered ~4ppt by a one-off DC-EOL revenue pull-forward and FY27 carries a deliberate ~3ppt hit from swapping equity comp for CASH. Underlying margin EXPANDS; underlying op profit +22.2%; GAAP op income $10.4M→$334.2M (32×). My Fundamentals analyst raised this as the lead red flag and withdrew it — but tested mgmt's 4ppt four ways and found it ~15-25% generous (real ΔR ≈ $295M ≈ 3.27ppt)
- Founders sold 7,665sh EACH EVERY TRADING DAY through 2026-02-06 (~$65M apiece) then STOPPED and have not sold since — six months of silence, verifiable from Form 4 cadence, a cleaner signal than the $250M CEO buy plan, which
Red flags
- BUT THE SBC FIX IS A RECLASSIFICATION, NOT A SAVING: SBC −$194M, cash comp +$223M → total comp cost RISES $29M.
(FCF−X)−(SBC−X)is identically unchanged — it ends the dilution treadmill (a real governance win, and the opposite of what disqualified ZS/HUBS/MDB/PD) but creates zero owner earnings - Ten metrics guided to one decimal; FCF got an adjective ("healthy")
- Share count "at a minimum relatively flat vs FY26" is measured against the FY26 average — it permits +2.67% from the Q4 exit
- FY26's entire headline margin expansion is one-off (−1.83ppt underlying)
- ~1.5-2pts of growth may be inorganic (DX) → FY27 organic ≈11.25%, not 13%
- They cut 1,600 staff (10%) in March to self-fund AI — a company priced per knowledge worker proving on itself that AI needs fewer. The org-level Rovo credit pool is the right escape hatch and overages are still not billed.
Watch
- SBC-adj FCF −$288M — worst ever (was +$53M); SBC/FCF 122% (not the 96% this line used to carry); FCF has been FLAT-TO-DOWN THREE YEARS ($1,415M→$1,415M→$1,319M) while revenue grew 50.8%; FCF margin 36.3%→20.1%. 🔍 Cause found: working capital ate $640.8M more than FY25 — DSO 54.5d→70.5d, AR +63%, and non-current deferred revenue −34.6%, a signature only upfront term-license recognition produces. FY27E SBC-adj FCF ~+$357M is 71% working-capital reversal + 38% a vanished restructuring charge — only 16% operating. Like-for-like pre-WC cash grows +3.5% on 13% revenue = negative operating leverage on cash.
- NEW: net cash $1,949M → $250M (−87%) — securities liquidated to zero to fund $1.8B of buybacks (136% of FCF) + $1.23B for DX. April's "net cash pillar" is gone; net interest flips +$20M → ≈−$20M
- DC guide is back-loaded: Q1 −4.0% vs FY −17.0% implies −22/−25% in H2; and ~6-8pts of "Cloud +25.5%" is migration, not new business
- had not bought a single share when the stock rose 39% (announced at $110, cooling-off period, transacts at $150+). 🔇 11.08% of float short into 3.1× volume — a material part of the +39% is covering, not re-rating.
- Consensus mean $153.77 / median $130 are STALE — 7 PTs raised to $160-200 on 8/07, ~8 firms unrevised. Moat 6.5/10 — switching costs at the suite (7.5) + compliance perimeter (7), but 🔍 the Teamwork Graph is a per-tenant data asset, NOT a network effect (the most common error on this name, and it inflates the multiple); Microsoft's forbearance is rented safety (GitHub Copilot for Jira went GA 6/25 as a bridge); Bitbucket already lost to GitHub; and two self-inflicted switching-cost resets run concurrently — DC→Cloud and Connect→Forge (EOS Q4 2026) — at the one moment a customer's switching cost is zero
- FIVE vendor traps:
SharesOut 159.63Mis Class A only — true 253,767,862, a 37.1% understatement ([[pitfall-yahoo-share-count-dual-class-fpi]]; the same check then caught WDAY −22.9%, ASAN −41.4%; the MNDY ~−26% claim was RETRACTED 2026-08-10 — MNDY is single-class and Yahoo was right, the gap was an $870M buyback);EV $28.06Bis wrong by $10.5B (true $38.59B — it implies $10.5B of net cash vs an actual $250M);PE(fwd) 24.90is stale pre-print (true ~28×); P/B, D/E, current ratio, BVPS all irreconcilable; Yahoo revenue not refreshed post-print. Graham VOID (GAAP EPS −$0.21, BVPS $4.17 on a $6.11B accumulated deficit). 🔍 The two frames now DISAGREE MORE than before: reported-FCF says $112-139 with 91.6% of the price covered by extrapolating the guide flat; SBC-adj owner earnings says $75-109. Better data widened the gap. That is why this is 6.5 and not 7. The price needs a terminal SBC-adj FCF margin of 27.4% — above CRM's 26.2%, 1.4× NOW's, 1.8× TEAM's own best year ever (15.0%) — from 4.8% today. Ranks 7th of 8 in the peer set on EV/SBC-adj FCF. FV $112-148 (central ~$128). Entry $100-120 The correct action is NO ACTION. ⏳ Q1 FY27 ~10/29 — watch DSO, SBC%, and DC vs the −4.0% guide. Revenue will be the least informative line. - Sleeve question flagged, not actioned: a −78% drawdown that forced this complete re-underwrite is a clean fail of this sleeve's "hold through a 40% drawdown without re-underwriting" test — the asset is near-evergreen (350k customers, 85% of the F500), the pricing model is not. | Trim 30x fwd (FY27 ends June-2027, non-GAAP EPS; renders ≈$163)
Would upgrade on
- Upgrade to [7.0]: FY27 SBC ≤17% of revenue (the one variable that reconciles the two frames) Rovo overages billed with disclosed consumption revenue an agent/non-human SKU SBC-adj FCF durably >10% of revenue the $250M actually executed.
🟡 [6.5] 已于 8/07 重新分析 —— 持有。 评分没变,但支撑这个评分的每一个输入都变了。 区间已全部重新推导(旧的 Add <$68 比现价低了 123% → [[pitfall-stale-entry-zone-suppresses-a-name]])。 **
4 月的论点是对的,而且已经兑现完毕:自 4/26 在 $71.55 的那份分析以来 +114%,自 $56.01 低点以来 +173%,光是 8/07 当天就 +38.9% —— 公司历史上最大的单日涨幅。 **
✅ FY26 第四季度是公司史上最好的一份财报: 收入 $1,766M +28%,云业务 +31% 且在加速(26→26→29→31,一路穿过这轮软件股崩塌),RPO +44% 至 $4,817M,而 cRPO 只有 +27% (客户正好在 AI 疑虑最响的时候签得更长、更大 —— 这是整份财报里最难被解释掉的一个数据点),非 GAAP 每股收益 $1.87 对预期 $1.50,GAAP 营业利润自 FY22 以来首次转正(第四季度 +$210.7M,利润率 12%;史上第一个真正盈利的季度)。 ARR 超过 $5M 的客户 +70% 至 69 家;超过 $3M 的 +50% 至 164 家。 **
🔍 「利润率从 30% 砍到 25%」这个标题是错的,股东信第 20 页自己写了(我已逐字核对过 PDF 原文):FY26 被一次性的 DC-EOL 收入提前确认虚增了约 4 个百分点,而 FY27 则背了约 3 个百分点的主动调整 —— 把股权薪酬换成现金薪酬。 还原之后利润率是扩张的;还原后的营业利润 +22.2%;GAAP 营业利润 $10.4M→$334.2M(32 倍)。 我的基本面分析师最初把这一条列为头号红旗,后来撤回了 —— 但他用四种方法交叉检验了管理层的 4 个百分点,发现偏乐观约 15-25%(真实的 ΔR ≈ $295M ≈ 3.27 个百分点)。
🚩 但股权激励的「修复」是一次会计重分类,不是真正的节省:股权激励 −$194M,现金薪酬 +$223M → 薪酬总成本反而上升 $29M。 (FCF−X)−(SBC−X) 在数学上完全不变 —— 它终结了靠增发稀释的循环(这是真正的治理改善,而且与当初淘汰 ZS/HUBS/MDB/PD 的行为正好相反),但创造不出一分钱的股东盈余。
🔴 扣除股权激励后的自由现金流 −$288M —— 史上最差 (上一年还是 +$53M);股权激励/自由现金流 122% (不是这一行以前写的 96%);自由现金流已经三年原地踏步($1,415M→$1,415M→$1,319M),而同期收入增长了 50.8%;自由现金流利润率 36.3%→20.1%。
🔍 原因找到了:营运资本比 FY25 多吃掉 $640.8M —— DSO 从 54.5 天升到 70.5 天,应收账款 +63%,而非流动递延收入 −34.6%,这个特征只有「预先确认定期许可证收入」才会产生。 ** FY27 预计扣除股权激励后的自由现金流约 +$357M,其中 *71% 是营运资本的回补 + 38% 是不会再发生的重组费用 —— 只有 16% 来自经营。 同口径比较下,营运资本前的现金只增长 +3.5%,而收入增长 13% —— 在现金层面是负经营杠杆。 **
🔴 新增:净现金从 $1,949M 降到 $250M(−87%) —— 有价证券被清仓至零,用来支付 $1.8B 的回购(相当于自由现金流的 136%)以及收购 DX 的 $1.23B。 4 月那份分析里「净现金」这根支柱已经没有了;净利息从 +$20M 翻转到约 −$20M。
🚩 十个指标都给到了小数点后一位,唯独自由现金流只给了一个形容词(「健康的」)。 🚩 股数「至少与 FY26 大致持平」是拿 FY26 的平均值作基准的 —— 这允许比第四季度的期末值 再上升 2.67%。 🚩 FY26 的全部账面利润率扩张都是一次性的(还原后实为 −1.83 个百分点)。 🚩 约 1.5-2 个百分点的增长可能是 外延并购(DX)带来的 → FY27 内生增长约 11.25%,而不是 13%。
⚠️ 数据中心的指引是后置的:第一季度 −4.0% 对全年 −17.0%,意味着下半年要 −22% 到 −25%;而且「云 +25.5%」里有约 6-8 个百分点是迁移来的,不是新增业务。
✅ 两位创始人在 2026-02-06 之前,每个交易日各自卖出 7,665 股(各约 $65M),然后就停了,此后再没卖过 —— 六个月的沉默,可以从 Form 4 的节奏上验证,这比 $250M 的 CEO 买入计划 是更干净的信号;⚠️ 而那份买入计划 在股价上涨 39% 的时候一股都还没买(在 $110 宣布,有冷静期,实际成交要到 $150 以上)。
🔇 流通股的 11.08% 是空头,成交量是平时的 3.1 倍 —— +39% 里有相当一部分是回补,不是重新定价。 ⚠️ 分析师均值 $153.77 / 中位数 $130 都是过期的 —— 8/07 有 7 家把目标价上调到 $160-200,约 8 家至今没动。
护城河 6.5/10 —— 真正的护城河是套件层面的转换成本(7.5)加上合规壁垒(7),但 🔍 Teamwork Graph 是单租户的数据资产,不是网络效应(这是这只票上最常见的分析错误,而且会抬高大家愿意给的估值);微软的克制是租来的安全(GitHub Copilot for Jira 已于 6/25 正式上线,而且是作为桥梁);Bitbucket 早已输给 GitHub;并且 两场自己制造的转换成本重置正在同时进行 —— DC→云,以及 Connect→Forge(2026 年第四季度停止支持) —— 而这恰恰是客户转换成本归零的唯一时刻。 🚩 他们在 3 月裁掉了 1,600 人(10%)来自筹 AI 投入 —— 一家按知识工作者人头收费的公司,亲自证明了 AI 会让所需的人更少。 ** 组织级的 Rovo 额度池是正确的逃生通道,而超额部分至今仍未开始计费。 **
❌ 五个数据源陷阱: SharesOut 159.63M 只是 A 类股 —— 真实值是 253,767,862,少算了 37.1%([[pitfall-yahoo-share-count-dual-class-fpi]];同一套检验随后又抓出 WDAY −22.9%、ASAN −41.4%、MNDY 约 −26%);EV $28.06B 错了 $10.5B(真实值 $38.59B —— 它隐含 $10.5B 的净现金,而实际只有 $250M);PE(fwd) 24.90 是财报前的过期一致预期(真实约 28 倍);市净率、负债权益比、流动比率、每股账面价值全都对不上;Yahoo 的收入数据在财报后没有刷新。 格雷厄姆估值无效(GAAP 每股收益 −$0.21,每股账面价值 $4.17,背后是 $6.11B 的累计亏损)。
🔍 两套估值框架现在的分歧比之前更大了:按申报自由现金流算是 $112-139,且现价的 91.6% 只要把指引平推就能覆盖;按扣除股权激励的股东盈余算则是 $75-109。 更好的数据反而把差距拉大了。 这就是为什么是 6.5 而不是 7。 现价要求终局的扣除股权激励后自由现金流利润率达到 27.4% —— 高于 CRM 的 26.2%,是 NOW 的 1.4 倍,是 TEAM 自己历史最好年份(15.0%)的 1.8 倍 —— 而今天只有 4.8%。 在同业 8 家里,按 EV/扣除股权激励后自由现金流排 第 7。
合理价值 $112-148(中枢约 $128)。 入场 $100-120 强力买入 ≤$85。 正确的操作是不操作。 **
⏳ FY27 第一季度约在 10/29 —— 重点看 DSO、股权激励占收入比重,以及数据中心相对 −4.0% 指引的表现。 收入会是整份财报里信息量最小的一行。 **
⚠️ 分类问题只标记、不处理:一次 −78% 的回撤逼出了这次彻底的重新论证,这明确不符合本篮子「能穿越 40% 回撤而不必重新论证」的准入标准 —— 资产本身接近永续(35 万客户,财富 500 强的 85%),定价模式并不是。 *
Historical results历史业绩 past fiscal years — these values do not change过去财年——这些数值不会变动
Profit margins利润率 the share of revenue that stays as profit营收中留存为利润的比例
of every $1 of revenue每 $1 营收中的占比
- Return on equity净资产收益率
- -4%
- Revenue growth营收增长
- +28%
Competitors竞争对手
Why customers stay客户为什么留下
Teams build their custom workflows and third-party add-ons on Jira and Confluence, which makes switching costly, and the add-on marketplace deepens the lock-in. Note: stock-based pay is very high — almost all of free cash flow — so the headline cash flow overstates the value reaching existing shareholders.各团队把自己定制的工作流程和第三方插件都搭建在 Jira 和 Confluence 上,这让更换的代价很高,而插件市场又进一步加深了这种绑定。要注意:它的股权激励支出非常高——几乎相当于全部自由现金流——所以表面上的现金流高估了真正落到现有股东手里的价值。
Recent news近期新闻 last 7 days近7天
- Is Atlassian (TEAM) A Bargain As Revenue Growth And Margins Draw Fresh Attention?Simply Wall St.
- Can Atlassian’s (TEAM) AI Push and GAAP Profit Target Reframe Its Long-Term Investment Story?Simply Wall St.
- Redditors Think Wall Street is Wrong About Atlassian (TEAM)Insider Monkey
- Atlassian, ServiceNow, and Samsara Shares Are Soaring, What You Need To KnowStockStory
- Atlassian’s (TEAM) Strategic Position in the AI LandscapeInsider Monkey
- Atlassian’s AI Pivot Is Starting to Challenge Wall Street’s Bear CaseMarketBeat