Cursor (Anysphere) — Series A Investment Memo
Thesis: Why this could be a fund-returner
Cursor is the first AI-native code editor to reach escape velocity among developers. It is not a thin wrapper around a large language model; it is a rethinking of the developer environment around AI. If Cursor becomes the default interface for AI-assisted software creation — the “operating system” for coding agents — it could capture a meaningful share of the tens of billions spent annually on developer tools and become a $10B+ company. At a hypothetical Series A, the entry valuation still leaves room for a 20x+ return if that outcome materializes. The speed of revenue ramp is the strongest signal I have seen in dev tools since GitHub.
Product & wedge
Cursor is a fork of Microsoft’s VS Code, which gives it an immediate distribution advantage: millions of developers already know the interface, and switching costs are near zero. But Cursor is not just VS Code with a chatbot bolted on. Its wedge is deep AI integration: tab autocomplete, codebase-aware chat, multi-file edits, and an agentic “Composer” that can plan and execute changes across a project. The product feels like a pair programmer that understands your entire repo, not a generic autocomplete.
The wedge works bottom-up. Individual developers experience a step-change in productivity within minutes, then bring it into teams. From there, Cursor expands into enterprise seats, security/compliance features, and eventually agentic workflows that go beyond editing into testing, deployment, and maintenance. The product is model-agnostic, routing to the best available LLM, which lets it stay ahead of any single model provider.
Market & competition
The market is large and growing: there are roughly 30 million professional developers worldwide, and global developer tool spend is estimated at $20B+ annually, with AI coding assistants the fastest-growing segment. Cursor competes directly with:
- GitHub Copilot — the default incumbent, bundled with GitHub and VS Code, with massive distribution.
- JetBrains AI — strong in enterprise Java/Kotlin ecosystems.
- Windsurf (Codeium) — an AI-native editor with similar ambitions and strong funding.
- Replit — cloud IDE targeting beginners and collaborative coding.
- Zed — performance-focused editor adding AI features.
- Sourcegraph Cody, Amazon Q Developer, Google Gemini Code Assist, Tabnine — all credible but less integrated.
Cursor’s differentiation is depth of AI integration and speed of iteration. It is not a plugin; it is the editor itself, rebuilt around AI. That is hard for incumbents to copy quickly without disrupting their existing products.
Traction & business signal
Publicly reported numbers are exceptional. Cursor went from roughly $4M ARR in early 2024 to a reported $100M+ ARR by late 2024, and reports in 2025 suggest $200M+ ARR — one of the fastest revenue ramps ever for a developer tool. Word-of-mouth among developers is unusually strong, and the product has become a status signal in engineering communities.
What is unknown from public information: net revenue retention, churn, gross margin after model inference costs, enterprise vs. individual mix, CAC payback, and long-term engagement depth. These are the metrics that will determine whether this is a durable business or a hype spike.
Risks — the three that actually kill the deal
- Microsoft platform risk. Cursor is a fork of VS Code, and GitHub Copilot is owned by Microsoft. Microsoft could bundle Copilot for free or near-free, restrict VS Code extension APIs, or copy Cursor’s agentic features and use its distribution to squeeze Cursor out. This is an existential threat. Cursor’s only defense is moving faster and building a brand that developers prefer, but platform owners have a history of killing popular forks.
- Model dependency and margin squeeze. Cursor relies on third-party LLMs (OpenAI, Anthropic, etc.). If model providers raise API prices, launch competing products, or withhold frontier models, Cursor’s gross margins could compress dramatically. Heavy usage of agentic features is expensive, and unit economics at scale are unproven. If inference costs eat most of revenue, the business may not be fundable at scale.
- Retention durability. Developer tools have low switching costs. Cursor’s growth may be driven by novelty and AI hype. If productivity gains plateau, or if a better model/UI emerges, users could churn quickly. Without strong net revenue retention and habit formation, the revenue curve could flatten or decline. Public data does not yet prove long-term retention.
Verdict: Invest at a hypothetical Series A
I would invest. The market timing, product quality, and revenue trajectory are too strong to ignore, and the entry price at Series A still allows for venture-scale returns. The risks are real, but they are the risks of a category-defining company.
The one question I need answered first: What is the 6-month net revenue retention for paying individual and team users? If it is above 120%, I would invest aggressively. If it is below 100%, I would pass — because that would mean the wedge is not sticky, and the growth is a leaky bucket.