Cursor outpaces expectations despite Microsoft competition, says a16z

If you needed a single number to understand how fast the AI coding tools market is moving, here it is: Cursor reached $100 million in annual recurring revenue faster than any software company on record. Not faster than most. Faster than any.

That milestone, flagged by venture firm Andreessen Horowitz in August 2026, came despite a competitive field that includes Microsoft’s GitHub Copilot and Anthropic’s Claude Code, two products backed by companies with effectively unlimited resources. Cursor’s CEO Michael Truell has made clear he’s not particularly worried about either of them.

From VS Code fork to $4B run rate

Cursor is built by Anysphere, a company founded by four MIT dropouts who initially worked on earlier AI projects before landing on an AI-native code editor as their core product. The editor is built on top of Microsoft’s open-source VS Code, which gave the team a head start on the interface side while they focused engineering energy on the AI layer underneath.

By mid-2026, Cursor’s annualized revenue had reached $4 billion, representing a 100% increase from 2023.

Andreessen Horowitz pointed specifically to Cursor’s model release cadence as a key driver of its outperformance. The company launched Composer 1.5 in February 2026, then followed up with Composer 2 just five weeks later. Both models achieved what the company describes as superior price-to-performance metrics compared to competing models from Anthropic and OpenAI, a claim that carries weight when your customers are developers who benchmark these things obsessively.

The product has also found its way into the largest companies in the world. Cursor now serves 60% of the Fortune 500, which suggests this is not a tool being adopted exclusively by startups and individual developers experimenting on weekends.

The iPhone moment thesis

Truell’s public framing of Cursor’s opportunity is worth paying attention to. He has described the current moment in software development as a precursor to multiple “iPhone moments,” a phrase that implies transformative shifts that change user behavior permanently rather than incrementally.

Cursor’s shift from a standard integrated development environment toward what it calls an agent-centric workspace reflects a broader thesis in the AI tooling space: that the next generation of developer tools won’t just autocomplete code but will actually execute multi-step tasks autonomously. The Composer model series is Anysphere’s attempt to build the engine for that kind of workflow.

What this means for the competitive landscape

Cursor’s growth trajectory puts real pressure on GitHub Copilot, which has the advantage of deep integration with GitHub’s existing infrastructure and Microsoft’s enterprise sales relationships. Copilot’s challenge is that it was built as an add-on layer rather than as the organizing principle of the editor itself. Cursor, by contrast, was designed from the start around the assumption that AI would be central, not supplementary.

Claude Code, Anthropic’s offering, represents a different kind of competition. It’s more narrowly focused on agentic coding tasks rather than serving as a full IDE replacement, which means Cursor and Claude Code aren’t necessarily competing for exactly the same user in every context. Truell’s stated comfort with that competitive dynamic suggests Anysphere sees the market as large enough to accommodate multiple winners, at least for now.

Disclosure: This article was edited by Editorial Team. For more information on how we create and review content, see our Editorial Policy.