Atlassian's 44% backlog jump is the number that answers the AI-agent bear case

Atlassian reported fourth-quarter revenue of $1.766B, up 28% year over year, and remaining performance obligations of $4.817B, up 44% ^[2][5]. The stock had traded around $84 in early July on the argument that AI agents would write the code, file the tickets and squeeze work-tracking tools out of the middle; it closed at $110.17 on the August 6 print and traded toward $146 the next morning, a 35% single-day move ^[7][9][8][1].

RPO is the line that speaks to the bear case, because it is contracted revenue not yet recognised. It grew at roughly twice the rate of subscription ARR, which came in at $6.606B, up 23% ^[5][4]. That is customers committing longer and larger, not trimming ahead of an agent-driven cleanout. SaaStr's reading is that in software development tooling so far, more agents has meant more work items, more branches, more reviews and more governance, and Atlassian sells the coordination layer for all of it ^[23].

Growth also went up rather than down at scale, which is the uncommon part. Cloud revenue grew 28% for the full fiscal year and 31% in the quarter; total revenue grew 26% for the year and 28% in the quarter ^[10][11]. Cloud is now about 69% of quarterly revenue ^[5]. GAAP operating income was $211M against a $28M loss a year earlier ^[6], across more than 350,000 customers including 85% of the Fortune 500 ^[14].

One deduction against the headline: reported revenue grew five points faster than subscription ARR ^[1]. SaaStr argues revenue is the lagging figure here, flattered by contract timing and Data Center renewals, and that 23% is the better read on the underlying run rate ^[19].

The cost of holding the category shows up in the guide, not the quarter. Atlassian is bundling its agentic capabilities into paid Jira Cloud at no additional charge, including orchestration of third-party coding agents such as Claude Code, Cursor and GitHub Copilot ^[15]. Rovo is used by more than 80% of the Fortune 500, and Rovo-assisted actions grew 50% quarter over quarter ^[16]. Non-GAAP operating margin hit a record 36% in the quarter, up from 24% ^[12], and the company then guided 28.5% for Q1 FY27 and 25% for the full year, with GAAP operating margin of 4.5% against 12% in the quarter ^[17]. That is 11 points of non-GAAP margin handed back ^[3], with non-GAAP gross margin trimmed from 88% in FY26 to a guided 86.5% ^[18]. FY26 free cash flow fell to $1.319B from $1.416B, a decline of about $97M ^[13][4].

Figma, which reported the day before, is the useful contrast. It grew 48% at a $1.5B run rate, a third consecutive quarter of acceleration, beat and raised, and the stock fell 15%, with attention landing on GAAP gross margin of 84%, down about five points year over year, because inference is a real per-request cost and some of it sits on beta products generating no revenue ^[20][21]. SaaStr's framing is that AI cost arrives in the P&L one to four quarters before AI revenue does ^[22]. Atlassian took it in operating margin and guidance; Figma took it in gross margin.

Two things to follow. Whether RPO growth holds near 44% or was pulled forward by long renewals, and whether subscription ARR converges up toward reported revenue or reported revenue decays back toward 23% ^[5][1][19]. And whether the 25% FY27 margin guide is sandbagging or the honest price of giving agentic features away free inside Jira ^[17][15].