Atlassian ships Secoda's data catalog for its AI agents; the test is holding ARR growth above 18.5%
Ten months after quietly buying the data-catalog startup Secoda, Atlassian has wired data warehouses into the graph its agents read, at a price no filing discloses. The bet: whoever supplies agents their context gets paid whichever model wins, and investors who cut the stock 17 percent last week want the revenue proof.
Richard Tangthree-minute read
Atlassian−4.88% — Atlassian, down 4.88 percent as of its last quote has turned December's acquisition of Secoda into shipped product, and no filing anywhere says what the deal cost. On October 7, at Team '26 Europe in Amsterdam, structured data became a native context type in the Teamwork Graph, Atlassian's index of an organization's work 13. Data lakes and warehouses now sit in that graph beside Jira tickets and Confluence pages, queryable from anywhere through Atlassian Insights. That is the Secoda catalog in production 13.
Secoda's catalog filled the graph's data hole
Secoda, a Toronto startup that catalogued and documented company data, announced it was joining Atlassian on December 4, 2025 2. TechTarget+6.76% — TechTarget, up 6.76 percent as of its last quote reported the purchase aimed at the structured-data gap in the Teamwork Graph, which knew who decided what in which ticket but not the numbers the business runs on 3.
The fill is live: the graph now connects more than 250 billion objects and relationships, and the new data context reads more than 20 sources, Databricks$218.01B — Databricks, private, latest valuation $218.01B, Snowflake+2.62% — Snowflake, up 2.62 percent as of its last quote and BigQuery among them 14. Unifying that index also concentrates what an intruder can reach; we reported today on attackers who crossed from a Confluence foothold to a cloud warehouse in four hours 5.
Outside agents already read it
The layer matters because it no longer stops at Atlassian's own apps. Atlassian rebuilt its server for the Model Context Protocol (MCP), the pipe through which outside AI tools read Jira, Confluence, Loom and Bitbucket 1. That server now carries more than 15 million tool calls a day, fifteen times its volume six months ago, and ranks among the most-used servers of Anthropic$350B — Anthropic, private, latest valuation $350B, the maker of Claude 6.
On November 5, Atlassian expanded its OpenAI$1.18T — OpenAI, private, latest valuation $1.18T partnership to ground ChatGPT and Codex in the same graph 6. The keynote's line, as one independent summary rendered it: you can rent models, but you cannot rent your company's work context 4.
The price never entered the record
EDGAR's full-text search, the SEC's filing database, finds Secoda in exactly two filings: Atlassian's fiscal 2026 annual report, which lists Secoda Inc. among its wholly-owned Canadian subsidiaries, and an unrelated 2005 loan document 78. No purchase price appears anywhere. A layer of the agent stack changed hands for a sum small enough to skip disclosure, and what Secoda's holders received stays between them and Atlassian.
The market wants the revenue hop
ServiceNow−2.41% — ServiceNow, down 2.41 percent today, the IT-operations software company, is building the same layer from the operations end, having announced its Context Engine in April on its Service Graph and Knowledge Graph, since fed by Veza's access graph and Armis's asset intelligence 910. Investors marked the gap between capability and cash on November 5: Atlassian beat estimates, guided fiscal 2027 subscription ARR growth down to about 18.5 percent, from 23 percent in the quarter, and the stock fell more than 17 percent 611.
The guide, not the benchmarks, settles the bet
Revenue growth has stayed above 14% for three years while the ARR guide falls to 18.5%
Data
| Revenue growth, year over year | |
|---|---|
| Q4 '23 | 21.5% |
| Q1 '24 | 29.9% |
| Q2 '24 | 20.5% |
| Q3 '24 | 21.5% |
| Q4 '24 | 21.4% |
| Q1 '25 | 14.1% |
| Q2 '25 | 22.3% |
| Q3 '25 | 20.6% |
| Q4 '25 | 23.3% |
| Q1 '26 | 31.7% |
| Q2 '26 | 27.6% |
| Q3 '26 | 23.7% |
In the results release, Mike Cannon-Brookes, Atlassian's CEO and Co-Founder, answered the question of what separates winners in enterprise AI: "Model power is table stakes." The advantage, he said, sits with platforms that give people and agents shared context to work from 6. The accuracy gains Atlassian claims for graph-grounded agents remain its own internal benchmarks 3. Turning the shipped layer into subscription growth is the remaining hop, and whether ARR growth holds above the 18.5 percent guide as the context layer ships is the number to watch 3.
Get Platforms by email, free
The cloud, data and developer software AI is built on. In our daily email, every weekday.






Reader comments
0 comments