On Wednesday morning, a PagerDuty alert fired inside Cursor's San Francisco office. A production incident, the kind that normally pulls an on-call engineer out of whatever they are working on, triggers a scramble through server logs, and eats up half a day. But no engineer responded. Instead, an AI agent spun up in a cloud sandbox, queried the logs through a Datadog connection, identified a recent code change as the likely culprit, drafted a fix, opened a pull request, and posted a summary in the team's Slack channel. The entire sequence took minutes.
This is Cursor Automations, the feature that Anysphere, Cursor's parent company, launched on March 5, 2026. And it arrives at a moment when the company's trajectory has become nearly impossible to overstate.
Three days earlier, Bloomberg reported that Cursor's annualized recurring revenue had doubled to $2 billion in just three months. For context, it took Slack five years to reach $1 billion in ARR. Zoom took nine. Cursor did the first billion in roughly three years and the second in a single quarter.
The Revenue Explosion No One Predicted
The numbers tell a story of compounding adoption that caught even bullish investors off guard. In late November 2025, when Anysphere closed a $2.3 billion Series D at a $29.3 billion valuation co-led by Accel and Coatue Management, the company reported crossing $1 billion in ARR. By February 2026, that figure had doubled.
The acceleration came primarily from enterprise customers. When Cursor hit $400 million ARR in late 2024, corporate buyers accounted for roughly 25% of revenue. By the $1 billion mark in November 2025, enterprise had grown to about 45%. Now, at $2 billion, nearly 60% of revenue flows from large corporate contracts, according to Bloomberg.
The pattern is consistent: companies that ran three-to-six-month pilot programs through mid-2025 began signing organization-wide deployments in Q4, locking in 500 to 5,000+ seats at $40 per month per developer on annual commitments. According to data from corporate spend platform Ramp, roughly 25% of generative AI clients now subscribe to Cursor in some capacity.
The user base has grown to match. Cursor now claims more than 7 million monthly active users, over 1 million daily active users, and more than 50,000 paying teams. The platform generates nearly one billion lines of code per day.
What Automations Actually Does
The pitch is straightforward: stop babysitting your AI agents. Until now, even Cursor's most advanced agent features required a developer to write a prompt, wait for the output, review it, and iterate. Automations breaks that loop entirely.
Engineers define a trigger and a set of instructions. When the trigger fires, an agent spins up in an isolated cloud sandbox, executes the task using whatever models and tool connections the team has configured, verifies its own output, and either delivers the result or loops in a human at a predefined checkpoint.
The triggers are the interesting part. Cursor has built native integrations with Slack, GitHub, Linear, and PagerDuty, plus a webhook system for custom events. A merged pull request can kick off a security audit. A new Linear ticket can spawn an agent that investigates the bug, checks for duplicates, and posts a root-cause summary. A cron job can generate a weekly digest of codebase changes and drop it into a Slack channel every Monday morning.
"It's not that humans are completely out of the picture," said Jonas Nelle, Cursor's engineering chief for asynchronous agents, in an interview with TechCrunch. "They're called in at the right points in this conveyor belt."
Josh Ma, an engineering lead at Cursor, added a detail that matters for understanding the product's ambition: "This idea of thinking harder, spending more tokens to find harder issues, has been really valuable." In other words, Automations does not just run the same agent logic on a schedule. It lets agents spend more compute time on deeper analysis because no one is sitting there watching a spinner.
Cursor says it already runs hundreds of automations per hour across its own codebase. The company's Bugbot system, which previously did basic code review, now performs full security audits and thorough architectural reviews on every pull request. The company says the system has "caught multiple vulnerabilities and critical bugs" that human reviewers missed.
Rippling and the Enterprise Playbook
Cursor's enterprise customers are not waiting around to experiment. Rippling, the HR and finance platform, has become one of the most aggressive adopters.
Engineers at Rippling have deployed automations across both personal and team workflows. One setup runs every two hours, aggregating meeting notes, action items, and TODOs from Slack, combining them with GitHub pull requests, Jira issues, and mentions, then producing a deduplicated dashboard for each developer. Other automations handle Jira ticket creation from Slack threads, incident triage, weekly status reports, and on-call handoffs.
This is the pattern Cursor is banking on: once one team inside a company sees the productivity gain, adoption spreads. Cursor's Bugbot alone now reviews more than 2 million pull requests per month across enterprise customers including Rippling, Discord, Samsara, Airtable, and Sierra AI.
The "Third Era" and the End of Prompting
To understand why Automations matters beyond the feature itself, you need to read what CEO Michael Truell published on Cursor's blog just a week before the launch.
In a February 26 post titled "The Third Era," Truell laid out a framework for how AI-assisted development is evolving. The first era was Tab autocomplete, the feature that made Cursor popular in the first place, handling keystroke-level tasks. The second era was synchronous agents: you write a prompt, the agent does work, you review and iterate. Truell says that era may last "less than a year" before the third takes over.
The third era is what Automations enables. Autonomous cloud agents that tackle larger, longer-horizon tasks with minimal human supervision. Truell shared a striking internal metric: 35% of the pull requests merged at Cursor are now created by agents operating autonomously in cloud virtual machines.
The developers who have adopted this workflow share three characteristics, according to Truell. They let agents write nearly 100% of their code. They spend their time breaking down problems and reviewing agent-produced artifacts rather than writing anything themselves. And they run multiple agents simultaneously rather than guiding a single one to completion.
Agent usage on Cursor has grown more than 15x over the past year. In March 2025, users of Cursor's Tab autocomplete outnumbered agent users 2.5 to 1. That ratio has now reversed: agent users outnumber Tab users 2 to 1.
"Cursor is no longer primarily about writing code," Truell wrote. The company's vision has shifted to building "the factory that creates software," with fleets of agents functioning as teammates.
This framing should concern every developer who has been treating AI coding assistants as sophisticated autocomplete. Cursor is not building a better text editor. It is building an autonomous workforce that happens to produce code.
The Competitive Arms Race Heats Up
Cursor is not the only company racing toward autonomous coding agents. GitHub Copilot shipped its own Agent Mode and now lets users assign Claude, Codex, or Copilot as the agent model on any issue. Anthropic's Claude Code operates as a command-line agent. OpenAI's Codex runs agents in sandboxed VMs. Windsurf's Cascade went fully agentic.
But Cursor holds a structural advantage: it controls the IDE. While Copilot works as a plugin inside VS Code and Claude Code runs in the terminal, Cursor is the editor itself, a fork of VS Code with AI woven into every layer. That means deeper context awareness, tighter integration with project structure, and now, with Automations, the ability to orchestrate background agents that interact with external tools through MCP (Model Context Protocol) connections.
The pricing tells its own story. GitHub Copilot Pro starts at $10 per month. Cursor Pro costs $20, with Pro+ at $60 and Ultra at $200. Cursor is charging more because it believes it can deliver more, and the revenue numbers suggest developers agree.
The question is whether that premium position holds as competitors close the gap. Every major player shipped cloud agent capabilities in February 2026. Microsoft has GitHub's distribution. Anthropic and OpenAI have the foundation models. Google has Gemini and the cloud infrastructure.
Cursor's bet is that execution speed matters more than platform size. The company went from $100 million to $2 billion in ARR in fourteen months. It shipped Background Agents, then Bugbot, then the Third Era blog post, then Automations, all in a span of weeks. That cadence is the moat, not any single feature.
What This Means for the People Who Write Code
The research on AI's effect on developer productivity has been mixed. A widely cited study found that AI tools actually made experienced developers 19% slower on certain tasks. But Cursor's internal data paints a different picture, one where the bottleneck is not the AI's capability but the developer's willingness to let go of the keyboard.
Truell's vision of the "third era" developer, someone who decomposes problems, reviews agent output, and runs multiple agents in parallel, sounds less like a software engineer and more like an engineering manager. That is either exciting or terrifying depending on where you sit.
For junior developers, the implications are particularly sharp. If agents handle the code and seniors handle the decomposition and review, where does a junior fit? Cursor's answer, implied if not stated, is that juniors become agent operators. They learn to specify tasks precisely, evaluate output critically, and orchestrate workflows rather than write for-loops.
For enterprises, the value proposition is clearer. At $40 per seat per month, an automation that catches one critical bug per quarter pays for itself many times over. The 60% enterprise revenue split suggests that this math is already winning budget battles inside large organizations.
The Bottom Line
Four MIT graduates built a code editor in 2022. Four years later, it generates $2 billion in annual revenue, runs hundreds of AI automations per hour, and has its CEO publicly declaring that the act of writing code by hand is entering its twilight. Cursor's Automations is not just a product launch. It is a thesis statement: the best software engineering teams of 2027 will not be the ones with the best programmers. They will be the ones with the best agents, and the best humans managing them.
The race to build always-on coding agents is no longer theoretical. It is running in production, triggered by Slack messages, and merging pull requests while you read this sentence.
Sources
- Cursor Recurring Revenue Doubles in Three Months to $2 Billion — Bloomberg (Mar 2, 2026)
- Cursor Is Rolling Out a New Kind of Agentic Coding Tool — TechCrunch (Mar 5, 2026)
- Cursor Has Reportedly Surpassed $2B in Annualized Revenue — TechCrunch (Mar 2, 2026)
- Automations: Always-On Agents — Cursor Blog (Mar 5, 2026)
- The Third Era of AI Software Development — Cursor Blog, Michael Truell (Feb 26, 2026)
- Cursor Revenue Doubles on Corporate AI Coding Demand — Investing.com/Bloomberg (Mar 2, 2026)
- Anysphere's Cursor Soars to $29B Valuation with $2.3B Round — TechFundingNews (Nov 2025)
- Cursor Automations Turns Code Review and Ops Into Background Tasks — Help Net Security (Mar 6, 2026)
- 35% of Merged PRs at Cursor Now Created by Autonomous Agents — OfficeChai (Feb 2026)
- Ramp AI Index February 2026 Update — Ramp (Feb 2026)
- Cursor's New Automations Launch Reimagines Agentic Coding — Dataconomy (Mar 6, 2026)