Cursor's last publicly reported revenue figure was $4 billion annualized in early June 2026. That sounds like annual revenue, but it is a run rate projected over twelve months. Forbes attributed the figure to a person familiar with the matter. It is not audited revenue for a completed fiscal year.
Other widely repeated Cursor figures need context too. The company does not publish a current number of paying individual customers. A JetBrains survey measures workplace use, not market share. And the $60 billion attached to the SpaceX acquisition belongs in a different category from a funding valuation.
The figures below include their dates, sources, and limits. The research cutoff is October 13, 2026.
- The latest public revenue figure is $4 billion annualized in early June 2026. It is a run rate, not audited annual revenue.
- Cursor's enterprise page says 50,000+ companies, 64% of the Fortune 500, 100M+ lines of enterprise code per day, and 93% preference in head-to-head evaluations. It does not give a method or measurement period.
- I could not verify a current public count of paying users. Cursor said millions of developers used its product in 2025, without giving a current paid-user or monthly-active-user count.
- Cursor raised Series A through D. The SpaceX acquisition, with an implied $60 billion equity value, became effective in August 2026 and was not a funding round.
- JetBrains' 18% and 12% figures describe workplace use in two different surveys, not market share.
- Cursor changed its plans on October 12, 2026. Monthly prices run from free through Plus ($8) and Premium ($30) to Ultra ($200); Teams starts at $30 per user.
1. What Cursor is today
Cursor is an AI development tool made by Anysphere. It launched in 2023 as a fork of Visual Studio Code. With Cursor 3, Anysphere introduced an agent-focused workspace in April 2026 that brings together multiple repositories, local and cloud agents, and a way to switch back to the IDE. This does not amount to a blanket promise that every VS Code extension or setting works unchanged.
In August 2026, Cursor began an early beta of Origin for paid plans, which includes code hosting, repository views, and pull requests with GitHub sync. Cursor's October 6, 2026 changelog also announced remote control of local agents through its iOS app. The host computer must remain on and online; the feature is off by default for Enterprise organizations.
Cursor's models and pricing documentation distinguishes Cursor Models from models offered by other providers. As of the research cutoff, the Cursor Models pool lists Grok 4.7, Grok 4.6, Grok 4.5, and Composer 2.5. The visible table for other models lists Claude Fable 5.1, Opus 5.5, and Sonnet 5.5, Gemini 3.1 Pro and 3.8 Flash, GPT-5.6 Luna, Sol, and Terra, plus Meta Muse Spark 1.3. The page hides additional models behind “Show more,” so this is not a complete catalog. Rates differ by model and mode. For example, Cursor lists Composer 2.5 Standard at $0.50 per million input tokens and $2.50 per million output tokens. Fast mode is $3 and $15, respectively. Cursor lists Claude Sonnet 5.5 at $2 input and $10 output per million tokens. These token rates are not the same as a monthly subscription price.
On August 28, 2026, OpenAI said it intended to end direct model access for Cursor after the SpaceX acquisition. OpenAI proposed November 12, 2026, as the date. The OpenAI Help Center says the final date has not been confirmed and Cursor could end access sooner. For local Chat and Agent requests, OpenAI lists a customer's own API key as an option. The Help Center says this does not apply to every Cursor feature, including Auto, Tab, and cloud agents.
2. User numbers and adoption
Cursor's current enterprise page gives four figures. Cursor does not publish a measurement date or detailed method there. I therefore label them as current company claims, not independently audited measurements.
Historically, Cursor said it had more than 40,000 “customers” in its August 2024 Series A post and that millions of programmers used Cursor as their editor in its January 2025 Series B post. The company did not define whether the 40,000 were paying or give an activity period. The millions figure also does not distinguish free from paying accounts or monthly active from registered users. Sources are Cursor, Series A and Cursor, Series B. As of the research cutoff, I could not verify a current number of paying individual customers or monthly active users in the available primary sources.
A country breakdown of cursor.com visitors would not be a breakdown of Cursor users. Website visits cannot be converted directly into customers, active users, or paid subscribers. I therefore do not present country shares as user statistics.
3. Funding and acquisition
The table separates capital raised from the company valuations reported at the time. Cursor confirmed the funding amounts for Series A, B, C, and D. The Series A and B valuations below come from contemporaneous TechCrunch reports. For Series B, its reported amount differs from Cursor's later announcement.
The seed round is supported by a contemporaneous TechCrunch report. For Series A through D, I use Cursor's posts on Series A, Series B, Series C, and Series D, plus TechCrunch's reports on Series A and Series B.
In April 2026, TechCrunch reported discussions about a possible further raise of more than $2 billion at a $50 billion pre-money valuation. It was not a completed Series E. The same report said Cursor forecast ending 2026 with an annualized revenue run rate above $6 billion. Both the funding discussion and the forecast were reported talks or projections, not a closed round or realized revenue figure. TechCrunch, April 17, 2026.
SpaceX agreed to the transaction on June 16, 2026, and completed it on August 14. The SEC filing gives an implied equity value of $60 billion for Cursor; consideration was primarily in SpaceX stock. Cursor is now a SpaceX subsidiary. This was not a Series E or another equity funding round, and its value is not directly comparable with a financing valuation.
4. Revenue milestones
Cursor has not used one revenue label consistently. The table preserves each source's wording. ARR, recurring revenue, and annualized run rate are not automatically the same metric. These are private-company figures, not audited annual revenue.
The earlier milestones come from Cursor's Series B, Series C, and Series D posts. For 2026, Bloomberg and Forbes reported the later figures.
TechCrunch reported in March 2026, citing Bloomberg, that large corporate buyers accounted for about 60% of Cursor's revenue. This is a dated estimate attributed to an unnamed source, not a company-published or audited segment breakdown.
Forbes also reported that Cursor said in late May that 75% of its run rate came from “other businesses.” Cursor did not further define that category in the quoted statement. The article also said the enterprise business grew threefold in the first quarter of 2026 compared with the fourth quarter of 2025.
The often repeated $6 billion figure is a year-end forecast reported in April 2026. It does not show that Cursor had $6 billion in October or has reached that run rate since. I found no later confirmed revenue figure by the cutoff date.
5. Workplace use in JetBrains surveys
JetBrains published two different surveys in 2026. The first was its AI Pulse Survey, with more than 10,000 professional developers worldwide. The second was its Developer Ecosystem Survey, with more than 15,000 respondents and statistical reweighting intended to better represent the global developer population. Their field periods, survey programs, and published methods differ. The sources do not identify the same respondents as a recurring panel.
The table preserves the published historical figures from both surveys. They are the shares of respondents who used each tool at work. The figures are not market shares or an exclusive distribution. The columns should not be read as one trend line. Sources are the JetBrains posts on its January AI Pulse Survey and May to July adoption trends.
JetBrains describes the change from 18% to 12% as a slight decline in workplace use. The figures come from different surveys; the published sources do not present them as a panel comparison of the same respondents. They measure neither paid subscriptions, revenue, nor market share.
6. Cursor pricing and model costs
Cursor changed its plans on October 12, 2026, as part of The X Subscription. As of October 13, 2026, the Cursor pricing page lists these monthly prices, excluding taxes and fees. All paid plans share one usage pool across Grok and Cursor. From Premium upward, third-party models are billed separately by usage and are not part of the monthly price. According to the announcement, existing subscribers on the earlier Pro, Pro+, and Ultra plans can stay or switch. Cursor's documentation still shows some of the old plan names at the cutoff.
A subscription price does not guarantee the same number of requests for every model. Cursor's documentation separates two usage pools. Cursor Models have different included limits from third-party models, whose token rates vary by model. Cursor bills Auto requests at the list price of the model that receives each request. According to the documentation (which still describes the earlier plans), third-party model requests on Teams and Enterprise also incur a Cursor Token Rate of $0.25 per million tokens, including requests routed through Auto. Usage therefore depends on the model and actual consumption.
For a model cost comparison, check the current models and usage pricing table before use. Plan, model, and mode all matter. The subscription alone is not enough to calculate actual usage or the cost of a particular request.
7. What the numbers do not prove
The public data is not enough to calculate Cursor's paid market penetration, the share of paying individual customers, or revenue by country. It also does not support claims such as “20 times faster than GitHub Copilot” or a reliable projection of when Cursor might overtake Copilot in paying customers. Comparable counts and time series are missing.
The company's historical posts document substantial growth. They do not establish superlatives such as the fastest-growing SaaS company of all time. That would require a defined comparison group and consistently measured revenue data. Run-rate figures are snapshots, not audited revenue.
The JetBrains surveys can help compare workplace use between Cursor and Claude Code, as long as their different methods remain clear. For more on Claude Code, see my article on Claude Code statistics. Other AI product figures are in my ChatGPT statistics.
