Anysphere closes $2.3B Series D for Cursor IDE at $29.3B valuation

On November 13, Anysphere — the maker of the AI-native code editor Cursor — closed a $2.3 billion Series D round at a post-money valuation of $29.3 billion, co-led by Coatue Management and Accel, with participation from existing investors Andreessen Horowitz, Thrive Capital, and DST Global, and new strategic partners Nvidia and Google, according to the company's announcement, a Business Wire press release, and reporting by CNBC and Bloomberg. The round closes with the company at over $1 billion in annualized recurring revenue.

Anysphere was founded in 2022 by four MIT alumni, with Michael Truell serving as chief executive. Cursor is built as a fork of Visual Studio Code rather than a plug-in, a structural choice the company has argued is necessary to make AI a first-class element of the editing surface rather than a side panel. The prior round, a Series C earlier in 2025, valued the company at roughly $9.9 billion; the November round therefore approximately triples the valuation in months, against an ARR profile that grew from roughly $200 million in March 2025 to over $1 billion by the close of the round.

In the company's "Past, Present, and Future" blog post and the accompanying Business Wire release, Truell wrote: "We believe that coding will be the single biggest driver of global productivity over the next decade, and our mission is to accelerate that progress. This funding allows us to dramatically increase our investment in research and product efforts and expand our footprint, ensuring we can continue to equip the world's engineering teams with the best tool for crafting software." Truell told CNBC in a same-day interview that the round reflects sustained product-market traction rather than a cyclical demand spike.

Anysphere reports more than 50,000 enterprise customers and a headcount of approximately 300 employees at the time of the announcement, according to the press release and reporting by CNBC and Bloomberg. The $1 billion ARR milestone places Cursor among the fastest software companies on record to reach the threshold; for comparison, GitHub Copilot — the incumbent in AI-assisted coding, owned by Microsoft and tightly coupled to the GitHub stack — had reportedly disclosed annualized revenue in the high hundreds of millions of dollars by mid-2025, while Codeium / Windsurf, the principal independent challenger, had been the subject of widely reported acquisition discussions.

Industry reaction was driven less by the round size than by the question of whether Cursor's plugin-versus-IDE bet has now been definitively answered. Several CTOs interviewed by Bloomberg argued that Cursor's IDE-fork architecture, combined with Composer-style multi-file agentic flows, has created enough product distance from Copilot to defend the valuation. Skeptical voices at The Information and the Financial Times noted that Cursor's gross-margin profile remains pressured by inference costs and that Anysphere's move to bring more inference in-house will be the more important number to track than headline ARR.

For us at Enpo Sekai, the Cursor round is a useful adjacent data point but not a roadmap input. Developer tooling is a market we observe, not one we plan to enter; our products live in characters, voice, and persona for desktop and games, and the Cursor business model — high-touch developer subscriptions priced against developer time savings — is structurally different from ours. What we do take from the round is a confirmation that AI-native, single-purpose desktop applications can scale into very large standalone businesses without becoming horizontal platforms, which is a pattern we are betting on at a much smaller scale.

We will be watching three things over the next twelve months: (1) Cursor's gross-margin trajectory once the in-house inference rollout is well underway, since that number determines how much of the topline ARR translates into durable operating leverage; (2) the competitive shape of the AI coding market after the Microsoft / Codeium / Windsurf situation resolves, and what that resolution implies for distribution dynamics across IDE forks and IDE plugins; (3) whether Cursor begins to expose its agent infrastructure — Composer, indexing, multi-file edits — to non-coding desktop workloads, which would change the addressable market and put it more directly in conversation with the persona-and-voice tooling layer we operate in.