Introduction: the cost of coordination
Asana sells software for a problem most companies do not measure. Employees have email, chat, video calls and shared drives, and still lose their day to finding files, sitting in status meetings, forwarding threads and chasing approvals. Asana's own research, cited in its Form 10-K for the fiscal year ended January 31, 2026, puts the figure at roughly 55% of working time spent on that busywork rather than on the work people were hired to do.
The company was founded in San Francisco in 2008 by Dustin Moskovitz, a co-founder of Facebook and its first chief technology officer, and Justin Rosenstein, an engineering lead at Facebook who had worked on Gmail and Google Drive at Google. Asana, Inc. (NYSE: ASAN) now has more than 180,000 paying customers, reported revenue of $790.8 million for fiscal 2026, and since mid-2026 has described itself as an operating system for human-agent teams rather than a project tracker.
From an internal Facebook tool to a company
The idea came out of Facebook's growth. As headcount climbed, no single person could say who owned what, and coordination slowed faster than the engineering organization grew. Moskovitz and Rosenstein wrote an internal tool called Tasks to fix it for themselves. It spread across the company without being mandated, which is the detail that mattered: if Facebook engineers adopted a project tracker voluntarily, the underlying problem was probably universal.
They left Facebook in 2008 and incorporated the business in Delaware on December 16 that year, initially under the name Smiley Abstractions, Inc., renamed Asana in July 2009 after the Sanskrit word for a seated yoga posture. Benchmark led a $9 million Series A in November 2009, with Andreessen Horowitz participating and Benchmark's Matt Cohler taking a board seat, at a point when there was no product to demonstrate.
Three years of building: the Work Graph
Most project tools of that era were either spreadsheets with extra features or rigid hierarchies inherited from desktop Gantt software. Both assume work is a list or a tree. Real work is neither: the same task matters to several teams at once, and each of them wants it in their own plan.
Asana's answer was the Work Graph, the proprietary data model its SEC filings describe as the institutional memory of the platform. Tasks, projects, portfolios, goals, people, files and discussions are stored as connected nodes, so a single item can live in several projects simultaneously without being copied. A task named "translate the app into Japanese" can sit in the marketing localization project, the engineering release sprint and the executive portfolio for the Tokyo launch. When an engineer changes its status, everyone reading a different view sees the same change, and nobody maintains a parallel copy.
Building that took time. The founders also wrote their own web framework, called Luna, to get the interface responsiveness they wanted. Asana opened to the public for free on November 2, 2011, roughly three years after incorporation, and only started charging in April 2012, when it also released a public API. Twitter and LinkedIn were among the first companies running work on it.
Funding and the 2020 direct listing
Asana raised in steps rather than continuously: $28 million in July 2012, $50 million in Series C funding in 2016 led by Sam Altman at a valuation of about $600 million, then $125 million across two 2018 rounds, including a $75 million Series D led by Generation Investment Management and a further $50 million that valued the company at $1.5 billion. Paying customers had passed 35,000 by January 2018.
On September 30, 2020 Asana went public through a direct listing rather than an initial public offering, so no new shares were sold and no underwriters set a price. The New York Stock Exchange published a $21 reference price, the stock opened at $27 and closed its first day at $28.80, valuing the company at about $5.5 billion. Asana added a second listing on the Long-Term Stock Exchange in August 2021.
Moskovitz then did something rare for a public-company founder: he bought more than $1 billion of Asana stock on the open market between June 2021 and early 2022, with his own money, and has continued to add smaller amounts since. Combined with a dual-class structure that gives Class B shares ten votes each, this left him holding a majority of the company's voting power, a position the fiscal 2026 10-K still describes.
Moving upmarket: Portfolios, Workload and Goals
Team-level task tracking does not close large contracts. Executives buying software for thousands of seats want to know how daily work connects to the outcomes they are accountable for. Asana built that layer in three releases. Portfolios arrived in November 2018 alongside a new Business plan, giving leaders one view of progress and resourcing across many projects. Workload followed in August 2019, showing how individual people were staffed so managers could rebalance before anyone was overloaded. Goals shipped on July 15, 2020, linking company objectives to the projects and tasks underneath them so progress updates as work completes rather than when someone writes a status report.
That sequence created the customer segments Asana now reports. At January 31, 2026 it had 25,928 Core customers spending $5,000 or more a year, contributing 76% of fourth-quarter revenue, and 817 customers spending more than $100,000. By July 31, 2026 those counts were 26,778 and 890.
The AI turn: Intelligence, AI Studio, AI Teammates and StackAI
Asana announced its first generative AI features in June 2023 and packaged them as Asana Intelligence that October: status summaries, action items pulled out of discussions, risk flags on slipping schedules. The argument was that large language models have reasoning but no organizational context, and the Work Graph is exactly that context.
In October 2024 the company shipped AI Studio, a no-code builder that lets a manager assemble an AI workflow for something repeatable, such as request intake and triage, and run it inside Asana. It exited fiscal 2026 with more than $6 million in annual recurring revenue and eight customers paying over $100,000 a year for it on top of their subscriptions. The energy retailer E.ON Next used it to automate project intake and triage, increasing the request volume it could handle by close to five times.
AI Teammates came next: agents that hold assignments inside a project and work from the same plan, context and governance as the people around them, rather than in a separate chat window. More than 200 customers joined the beta before general availability for sales-led customers in fiscal 2027. Both AI products are billed on consumption rather than per seat, and management expects AI offerings to contribute roughly 15% of new annual recurring revenue in fiscal 2027.
The missing piece was execution outside Asana, and the company bought it. On May 28, 2026 Asana completed the acquisition of StackAI, a no-code AI workflow platform, for $75 million, its first acquisition in 18 years; StackAI founders Tony Rosinol and Bernard Aceituno joined the company. StackAI lets agents act across the systems where enterprise data actually sits, including ERP, CRM and ITSM tools. In June 2026, at its Work Innovation Summit, Asana used that foundation to relaunch the platform as an operating system for human-agent teams.
A new chief executive
Moskovitz announced in March 2025 that he would retire as chief executive, and the stock fell 25% that day. In June 2025 Asana named Dan Rogers, then chief executive of LaunchDarkly and previously an executive at ServiceNow and Rubrik, and he took over on July 21, 2025 with Moskovitz becoming board chair. Rogers added the chair role himself effective September 25, 2026, when Moskovitz moved to a director seat and Jerry Ting and Tom Berquist joined the board.
The finance seat turned over twice in the same stretch. Tim Wan, chief financial officer from 2017 and through the direct listing, was succeeded by Sonalee Parekh in September 2024, who resigned effective March 23, 2026. Aziz Megji, previously head of financial planning and analysis, was promoted to the role the following day. Chief Operating Officer Anne Raimondi and General Counsel Eleanor Lacey both resigned effective December 31, 2025.
What the numbers say
Asana's fiscal year ends January 31. Revenue grew from $76.8 million in fiscal 2019 to $790.8 million in fiscal 2026, but the growth rate fell every year after fiscal 2022: 66.7%, then 44.6%, 19.2%, 10.9% and 9.2%. Profitability moved the other way. Gross profit reached $704.0 million in fiscal 2026, an 89% margin on cost of revenues of $86.8 million. Non-GAAP operating income turned positive at $56.7 million against a $40.8 million loss a year earlier, and adjusted free cash flow rose to $84.5 million from $2.6 million. The GAAP net loss was $189.0 million, down 26% year over year, still mostly stock-based compensation. Headcount fell 3% to 1,767 employees. Cash, cash equivalents and marketable securities stood at $434.0 million.
Revenue splits roughly 59% United States and 41% international, at $468.3 million and $322.5 million in fiscal 2026. Second-quarter fiscal 2027 revenue, reported on September 3, 2026, was $216.4 million, up 10%, with non-GAAP operating margin near 10% and full-year guidance raised.
The open question
The hard number in Asana's disclosures is retention. Dollar-based net retention was 97% for Core customers and 96% for accounts spending over $100,000 at January 31, 2026, which means the installed base shrank slightly and all net growth came from new customers. Management has also said that AI-driven changes in search and paid media cut the self-serve top of the funnel, with roughly a two-point drag on annual recurring revenue growth assumed through fiscal 2027 and no recovery built into guidance.
So the bet is legible. Asana is wagering that the Work Graph, built over three years for human teams, is the record that AI agents need in order to be trusted with real work, and that customers will pay for agent output on consumption even as they buy fewer seats. The products exist, the margin structure now supports the experiment, and the next few quarters of net retention will show whether the wager is paying.
