Anysphere
AI-native code editor — the maker of Cursor, a VS Code-fork IDE with multi-model routing across Claude, GPT-5/Codex and Gemini plus an in-house Composer model, scaling from $100M to $2B annualised run-rate revenue in roughly 13 months on a ~~180-person team (press-triangulated). SpaceX announced a $60B acquisition agreement on 22 Apr 2026 (TechCrunch); the pending $2B Series E was halted as a binding acquisition agreement is now in place (Bloomberg, 19 May 2026).
The Business
Anysphere is the maker of Cursor — an AI-native code editor built as a fork of Microsoft’s open-source VS Code. The product layers in-line AI completion, multi-turn chat with the active codebase, an agent surface (“Cloud Agents”) and a tab-completion model on top of the VS Code foundation. The model layer routes across Anthropic Claude Sonnet/Opus, OpenAI GPT-5/Codex models and Google Gemini, with Anysphere’s in-house Composer 2.5 model (shipped November 2025) added to dilute third-party model dependence and improve unit economics. Distribution is direct-download for individuals and self-serve for small teams, with an enterprise tier for Fortune 1000 deployments. The company was founded in 2022 by four MIT classmates — Michael Truell (CEO), Sualeh Asif (COO), Aman Sanger and Arvid Lunnemark; Lunnemark departed in October 2025, leaving the remaining founder trio in place.
Customers and Distribution
Anysphere disclosed approximately 60% enterprise revenue mix and 70%+ Fortune 1000 active customer penetration at the April 2026 Series E talks reporting point, with SpaceX named in evaluation as an anchor enterprise customer. Annualised run-rate revenue trajectory: $100M in January 2025, $500M in June 2025, $1B in November 2025, $2B in February 2026 — the fastest B2B path to $1B and $2B ARR on the public record. Internal projection of $6B ARR by year-end 2026 per the Fortune founder profile and TechCrunch coverage. Distribution sits across two surfaces: direct-download Cursor desktop install across macOS, Windows and Linux (with VS Code extension and keybinding import on first launch as the inbound-migration accelerant), and the Anysphere business / enterprise tiers for self-serve and procurement-driven enterprise deployment. The ~50-employee team at $2B+ annualised run-rate revenue implies revenue-per-employee in the order of $40M — the highest in the IM tracker universe and an order of magnitude above the enterprise-software median.
Model Strategy
Anysphere is a coding-specific vertical-AI play with multi-model routing as the operating posture and an in-house Composer model as the strategic hedge. The routing layer covers Anthropic Claude Sonnet/Opus, OpenAI GPT-5/Codex, Google Gemini 2.5 Pro and Composer 2.5 internally; the company positions Composer as the lever to “reduce reliance on third-party AI models and improve unit economics,” shipped November 2025 in response to the supplier-concentration risk that all three of Cursor’s largest model suppliers are also direct first-party coding-agent competitors. The Cursor IDE itself is the product moat — tab-completion latency and quality, multi-turn chat with codebase context, the Cloud Agents API surface, and the VS Code-fork architecture that lowers inbound switching cost from VS Code while raising outbound switching cost to alternatives (no official chat-history exporter; Cloud Agents API is not OpenAI-compatible). The $2B Series E talks at a reported $50B pre-money (post-money ~$52B), co-led by Andreessen Horowitz and Thrive with NVIDIA as a strategic co-investor, were halted in April 2026 when SpaceX announced a $60B acquisition agreement (TechCrunch 22 Apr 2026; Bloomberg 19 May 2026 confirmed the buyout terms with close conditioned on SpaceX’s IPO timing).
At A Glance
The Numbers
Annualised revenue
Paid subscribers
Headcount (FTE)
Funding to date
Leadership Team
Anysphere is unusually founder-led for a company at $2B+ annualised run-rate revenue. No CRO, COO (separate from co-founder Sualeh Asif), CFO or CTO has been publicly disclosed as a separate senior appointment; the ~50-person organisation is highly leveraged, with revenue-per-employee approaching $40M, but the bench depth at the senior level is thin relative to peers. The Fortune founder profile (April 2026) flags the founder-led GTM organisation as a structural watch-item. The pending SpaceX acquisition agreement (announced 22 Apr 2026; Bloomberg confirmation 19 May 2026) raises the prospective question of leadership integration into SpaceX rather than a CRO/COO/CFO bench build at Anysphere itself.
IM Framework Scoring
IM’s structured assessment of Anysphere’s competitive position. The summary below is the headline; expand “Show the full analyst-grade analysis” near the bottom for the per-dimension reasoning and evidence. Methodology →
Funding History
| Date | Round | Raised | Post-money | Lead investor(s) |
|---|---|---|---|---|
| Apr 2026 | SpaceX acquisition agreement (pending close) | ~$60B buyout | ~$60B headline buyout value | SpaceX (announced 22 Apr 2026; Bloomberg 19 May 2026) |
| 2026 | Series E (halted) | ~$2B (talks) | ~$52B (reported $50B pre-money) | Andreessen Horowitz & Thrive co-led talks; NVIDIA strategic co-invest — round halted by SpaceX acquisition |
| Mid 2025 | Series C | ~$900M | ~$9.9B | Thrive Capital, Andreessen Horowitz, Accel |
| Nov 2025 | Series D | ~$2.3B | ~$29.3B | Accel & Coatue co-led (NVIDIA and Google new investors) |
| Dec 2024 | Series B | ~$105M | ~$2.6B | Andreessen Horowitz (Thrive, Benchmark, OpenAI participating) |
| 2024 | Series A | ~$60M | ~$400M | Andreessen Horowitz |
| 2023 | Seed | ~$8M | — | OpenAI Startup Fund |
Cumulative external equity well above $3B before the SpaceX acquisition agreement. On 22 Apr 2026 SpaceX announced a $60B acquisition agreement (TechCrunch) preempting Anysphere’s then-pending $2B Series E by offering a $60B buyout to close 30 days post SpaceX IPO; Bloomberg confirmed the binding buyout terms on 19 May 2026. The Series E talks at a reported $50B pre-money (post-money ~$52B), co-led by Andreessen Horowitz and Thrive with NVIDIA as a strategic co-investor, were halted. The mid-2025 Series C at ~$9B was led by Thrive with Andreessen Horowitz and Accel participating — the round at which Anysphere first crossed $500M annualised run-rate revenue. Series A and B figures and dates triangulate from Anysphere’s own blog, TechCrunch and Forbes coverage and may carry +/- 1-month dating tolerances pending a primary disclosure.
Competitive Landscape
Anysphere’s competitive set sits in three concentric rings: foundation-model providers shipping first-party coding agents on the models that Cursor itself routes to (Anthropic Claude Code, OpenAI Codex), the incumbent enterprise developer-tools channel built on top of the GitHub and VS Code distribution surfaces (Microsoft GitHub Copilot, plus the Codeium / Windsurf play that Microsoft / OpenAI moved on in early 2025), and the long-standing professional-IDE vendors and browser-native challengers (JetBrains, Replit). The asymmetry that defines the set is that Anthropic, OpenAI and Google are simultaneously Cursor’s largest suppliers and its three most credible competitors at the agent-coding surface.
| Competitor | Positioning | Distribution edge | Threat profile |
|---|---|---|---|
| GitHub Copilot (Microsoft (NASDAQ: MSFT)) |
The incumbent AI coding assistant — first to market, deepest GitHub-repository integration, and the most-installed enterprise developer-tool in the cohort. Now offers multi-model routing across OpenAI, Anthropic and Google. | GitHub repository footprint, Microsoft 365 Copilot bundle, VS Code default placement, GitHub Enterprise Server. | High — channel control through GitHub and VS Code is structural; Copilot is the default option for every enterprise that procures Microsoft 365. |
| Claude Code (Anthropic) |
Anthropic’s first-party developer-coding agent and CLI — bundled with Claude API subscriptions, with rapid feature cadence and Claude Opus / Sonnet model access at the source. | Direct via Anthropic API and Claude.ai, plus AWS Bedrock distribution alongside Claude on Bedrock. | High and asymmetric — Anthropic is Cursor’s single largest model supplier on the routing layer and a direct competitor at the coding-agent surface. |
| OpenAI Codex (OpenAI) |
OpenAI’s re-launched developer-coding agent — bundled into ChatGPT subscriptions and the OpenAI API. Codex CLI and Codex web together are OpenAI’s bet at the same surface Cursor occupies. | ChatGPT consumer footprint (900M+ weekly active users on the March 2026 funding-round disclosure), OpenAI API, Microsoft Foundry / Azure OpenAI partnership. | High and asymmetric — same supplier-vs-competitor dynamic as Anthropic, with consumer-distribution advantage at the front of the funnel. |
| Replit | Browser-native coding environment plus AI agents for prototyping and lightweight production work. Higher up the abstraction stack than Cursor; targets a different developer-persona but converges in the small-team and prosumer segment. | Browser-only direct; large education and prosumer footprint; emerging enterprise tier. | Medium — less head-to-head on the desktop-IDE professional-developer use case where Cursor leads, more competitive on agent-led code generation for non-traditional developers. |
| JetBrains AI Assistant (JetBrains) |
AI capability layered into the JetBrains family of professional IDEs (IntelliJ, PyCharm, WebStorm, GoLand) — the long-incumbent paid-IDE alternative to VS Code with the deepest static-analysis and refactoring tooling. | JetBrains IDE install base (~9M paid developers per company disclosure), JetBrains Toolbox subscription channel. | Medium — not VS Code-fork architecture so portability friction runs both ways, but JetBrains owns a large pre-existing paid-IDE relationship that Cursor must continue to displace. |
Pricing benchmark: Cursor consumer pricing (Free / Pro $20 / Business $40 / Enterprise custom) tracks the GitHub Copilot consumer band (Copilot Individual $10 / Business $19 / Enterprise $39). Claude Code is bundled into Claude API and Claude.ai Pro/Max subscriptions; OpenAI Codex is bundled into ChatGPT Pro and the OpenAI API; JetBrains AI Pro sits at $19/month on top of an existing JetBrains subscription. The competitive frame is capability cadence, model-routing flexibility and developer-mindshare rather than headline per-seat price — on which Cursor is broadly at parity with Copilot.
Potential Risks
The case for Anysphere at IM Framework 7.46 rests on the fastest B2B revenue ramp on the public record ($100M to $2B ARR in ~13 months), the ~~180-person team (press-triangulated) at $2B+ ARR (~$40M revenue per employee), the pending Series E at a reported $50B pre-money (post-money ~$52B) with NVIDIA strategic co-investment, the ~60% enterprise revenue mix with 70%+ Fortune 1000 active customer penetration, and Composer 2.5 as the in-house model bet. The case against splits into five risks of differing magnitude — with foundation-model supplier-and-rival concentration the most structural and EU AI Act deployer-obligation exposure the most date-certain.
Foundation-model supplier-and-rival concentration
Anysphere’s headline model routes are Anthropic Claude Sonnet/Opus, OpenAI GPT-5/Codex models and Google Gemini — all three of which ship first-party coding agents (Claude Code, OpenAI Codex, Gemini Code Assist) that compete head-to-head with Cursor at the developer-IDE surface. The in-house Composer 2.5 model (shipped November 2025, positioned explicitly to “reduce reliance on third-party AI models and improve unit economics”) is the strategic response, but at the current Composer share of Cursor inference the structural dependency on the three model rivals remains material. The sub-rubric score on D4a supplier diversity is 6 on this evidence.
EU AI Act deployer-obligation exposure from August 2 2026
The EU AI Act high-risk and general-purpose AI obligations (Articles 9-17, Article 26 deployer obligations, Article 50 transparency) come into binding force on August 2 2026. Cursor was assumed low-regulation as a coding tool but enterprise-deployer obligations — risk-management documentation, log retention, conformity assessment for high-risk use cases — bring real compliance burden into regulated-customer procurement. Cursor 3 has been flagged in third-party analyses as harder to deploy in regulated contexts than Claude Code or Codex, with documentation gaps for regulated environments. The sub-rubric score on D4c regulatory exposure was downgraded from 8 to 6 on this evidence.
Key-person dependency on a ~180+ employee bench (press-triangulated) at $2B+ ARR scale
Anysphere runs a ~50-employee organisation at $2B annualised run-rate revenue — revenue-per-employee in the order of $40M, an extraordinary capital-efficiency outcome, but a bench-depth profile that creates concentrated key-person risk. Co-founder Arvid Lunnemark departed in October 2025; coding researchers Boris Cherny and Cat Wu were poached back by Anthropic mid-2025. The remaining founder trio (Truell, Asif, Sanger) carries the senior weight. No CRO, COO (separate from Asif), CFO or CTO has been publicly disclosed as a separate senior appointment. The sub-rubric score on D4e is held at 4.
Customer portability runs asymmetrically against Cursor
Cursor’s VS Code-fork architecture is a portability strength on the inbound side (one-click VS Code extension and keybinding import) but a structural weakness on the outbound side: Cursor-to-VS-Code migration requires manual chat-history export with no official importer, and the Cloud Agents API is non-OpenAI-compatible (not drop-in-substitutable into rival agent stacks). Composer 2.5 creates a new in-house-model lock-in dimension on Cursor users on its routes. The sub-rubric score on D1c portability was downgraded from 5 to 4 on this evidence: substitution friction is real and growing.
SpaceX acquisition close and regulatory clearance
SpaceX announced a $60B acquisition agreement on 22 Apr 2026 (TechCrunch) preempting the pending Series E; Bloomberg confirmed the binding buyout terms on 19 May 2026 with close conditioned on the SpaceX IPO timing (30 days post-IPO). The principal risks at this stage are antitrust review of a SpaceX coding-tool combination, SpaceX-IPO timing slippage extending the close window, integration friction between Cursor’s founder-led organisation and SpaceX’s operating model, and customer churn during the close window if competitive coding-tool incumbents accelerate displacement attempts.
Valuation-to-revenue gap at the reported $50B Series E
A reported ~$50B pre-money valuation (post-money ~$52B) on $2B annualised run-rate revenue is a ~25x multiple. The growth trajectory ($100M ARR January 2025 → $500M June 2025 → $1B November 2025 → $2B February 2026, projecting $6B by year-end 2026 per internal projection) is the fastest B2B revenue ramp on the public record and justifies a premium. The compression risk is real if Composer in-house model share does not scale fast enough to dilute Anthropic/OpenAI inference dependence, if Claude Code or OpenAI Codex compresses Cursor’s developer-preference lead, or if the EU AI Act deployer obligations slow enterprise procurement in Europe.
Recent IM Coverage
- AI Coding Agents — Category Report May 2026.
- Coding Agents — the Margin Reset is Coming Apr 2026.
Show recent press coverage of Anysphere
- May 2026 — SpaceX is said to plan to buy startup Cursor 30 days after IPO — $60B acquisition agreement confirmed.
- Apr 2026 — SpaceX announces $60B acquisition agreement for Cursor (Anysphere), preempting Series E.
- Apr 2026 — Cursor in talks to raise $2B Series E at $50B valuation as enterprise growth surges.
- Apr 2026 — Cursor: who is Michael Truell, the 25-year-old CEO building the fastest-growing B2B startup on record.
- Apr 2026 — Anysphere (Cursor) lines up $2B at $50B valuation.
- Mar 2026 — Cursor at a crossroads: Anthropic supplier-vs-rival dynamics inside the fastest-growing developer tool.
- Nov 2025 — Cursor ships Composer 2.5 in-house model to reduce reliance on third-party AI models and improve unit economics.
- Jun 2025 — Cursor crosses $500M ARR on Series C at ~$9B valuation — Thrive Capital leads.
- 2024 — Anysphere raises $60M Series A led by Andreessen Horowitz; Cursor positioned as the AI-native VS Code fork.
Curated feed of named-source coverage — Anysphere’s own blog posts, named-author TechCrunch, Fortune, Forbes and The Next Web coverage. Excludes paywalled article bodies of The Information, WSJ, FT and Bloomberg (headline + free-snippet only), PR-wire reposts of the same release and unsourced AI round-up pieces. The Forbes 2024 Series A reference carries +/- date tolerance pending a primary Anysphere disclosure.
Show the source register for the figures on this page
IM operates a primary-source-where-possible discipline. The figures above come from:
- Revenue: Anysphere is private and does not file. Annualised run-rate revenue crossed $2B in February 2026 per The Next Web and the TechCrunch Series E talks, up from $1B in November 2025, $500M in June 2025 and $100M in January 2025 — the fastest B2B path to $1B and $2B ARR on the public record. Internal projection of $6B ARR by end of 2026 per the same coverage. We label this “annualised run-rate revenue” rather than “GAAP revenue” because Anysphere does not publish GAAP financials.
- Usage — enterprise penetration: Per the Fortune founder profile and the TechCrunch Series E talks: approximately 60% of revenue is enterprise; more than 70% of Fortune 1000 companies are active customers; SpaceX is named in evaluation as an anchor enterprise. We decline-to-publish a precise paid-seat figure pending a primary Anysphere disclosure.
- Headcount (basis-disclosure note): Anysphere does not publish precise headcount. Public references to the team size cluster around ~180+ employees (Wikipedia-triangulated; precise headcount not company-disclosed) at the $2B-ARR disclosure point per Wikipedia (Cursor company entry) and the Fortune profile. We carry “~50” as an order-of-magnitude reference rather than a precise figure pending a primary Anysphere disclosure.
- Funding to date: Cumulative external equity well above $3B excluding the pending Series E. References: TechCrunch Series E talks at $50B pre-money (post-money ~$52B); The Next Web — $2B at $50B with NVIDIA strategic co-invest; mid-2025 Series C at ~$9B led by Thrive with Andreessen Horowitz and Accel; early-2025 Series B at ~$2.5B led by Thrive; 2024 Series A at ~$400M led by Andreessen Horowitz; 2023 seed by OpenAI Startup Fund.
Methodology & Disclaimer
For metric definitions, source-tier hierarchy, and decline-to-publish rules, see the tracker methodology. Confidence dots (• green / • amber / • red) follow the same convention as the AI Tracker.
Spotted a figure you believe is wrong? Send corrections to info@informationmatters.net.
Information Matters Framework scores are the considered opinion of the IM team — human and AI — applied to publicly-available evidence under a disclosed methodology. They are not statements of fact about the companies scored and they are not investment advice.
