Cursor hit $1 billion in annual recurring revenue in November 2025. That’s just 24 months after the company launched.
This makes Cursor the fastest-growing B2B software company in history. The AI code editor beat the growth rates of Slack, Zoom, and every other enterprise software company before it.
But here’s the catch: Multiple reports say Cursor spends 100% of its revenue on AI costs. That means despite making a billion dollars, the company isn’t profitable yet.
This breakdown shows how Cursor makes money, who pays them, and whether the business model can actually work long-term.
The Numbers: Cursor Revenue Growth
The Cursor revenue story is remarkable. Here’s how fast the company grew:
- January 2025: $100M ARR
- April 2025: $300M ARR
- June 2025: $500M ARR
- November 2025: $1B ARR
In the first half of 2025, revenue was doubling approximately every two months. That’s unprecedented growth for enterprise software.
The company’s valuation followed the same trajectory. Cursor raised $2.3 billion in Series D funding at a $29.3 billion valuation in November 2025. Total funding now stands at $3.3 billion from investors including Accel, Coatue, Thrive Capital, Andreessen Horowitz, Google, and NVIDIA.
How Cursor Makes Money
Cursor generates revenue by selling subscriptions to individual developers and companies. The business model breaks down into three main segments.
Individual Developers Generate $370M
The Pro plan costs $20 per month and serves the core developer market. The Ultra plan costs $200 per month for power users who need more compute resources.
Over 1 million developers now pay for Cursor individually. Most use the Pro tier at $20 monthly, while power users who burn through AI credits faster upgrade to Ultra.
These individual users contribute about 37% of total revenue, or roughly $370 million annually.
Who uses Cursor individually? Mostly independent developers, freelancers building projects, and programmers working on side businesses. The product appeals to developers because it’s built on VS Code, which means no learning curve, and the AI autocomplete actually works well.
Teams Generate $336M
Companies buying Cursor for entire engineering teams pay $40 per user per month. This tier includes features businesses need like single sign-on, admin controls, and centralized billing.
Major tech companies use Cursor for their engineering teams. The customer list includes Shopify, OpenAI (yes, OpenAI uses Cursor), Midjourney, Stripe, Uber, Adobe, and Spotify.
With approximately 700,000 team seats deployed, this segment generates about $336 million in annual revenue, or 34% of the total.
Enterprise Accounts Generate $290M
The Enterprise tier uses custom pricing, typically ranging from $500 to $1,000 per user annually. These deals include compliance features, audit logs, pooled usage credits, and dedicated support that large companies require.
More than 50% of Fortune 500 companies now use Cursor. Enterprise customers include NVIDIA, whose CEO Jensen Huang called Cursor his “favorite enterprise AI service.” Google uses Cursor despite being an investor. Microsoft teams use it despite Microsoft owning GitHub Copilot.
This enterprise segment contributes about $290 million annually, representing 29% of total revenue.
What You Get at Each Cursor Pricing Tier
Understanding what each plan offers helps explain how Cursor converts users from free to paid.
The Free plan gives unlimited basic Tab autocomplete plus 50 premium model requests monthly. This lets developers try the product without paying anything.
The Pro plan at $20 monthly includes unlimited Tab autocomplete, unlimited Auto mode where Cursor picks the AI model, and a $20 credit pool for premium models. The credits work out to roughly 225 Claude requests or 550 Gemini requests, depending which AI model you choose.
The Ultra plan at $200 monthly provides 20 times the Pro usage with no compute caps. This targets developers who use Cursor constantly throughout their workday.
The Teams plan at $40 per user monthly adds collaboration features. Companies get SSO, admin controls, centralized billing, and usage analytics showing how each team member uses the product.
The Enterprise plan uses custom pricing and adds SCIM provisioning, audit logs, pooled credits, dedicated support, and invoice billing. Large companies need these features for compliance and security.
The Profitability Problem
Despite generating $1 billion in revenue, Cursor isn’t profitable. Multiple sources report the company spends approximately 100% of revenue on AI model costs from Anthropic and OpenAI.
The math looks like this:
- Revenue: $1 billion
- AI API costs: ~$1 billion (100% of revenue)
- Employee costs: ~$75 million (300 employees)
- Infrastructure: ~$25 million
- Other expenses: ~$50 million
Total costs run about $1.15 billion against $1 billion revenue, creating a $150 million annual loss.
This means Cursor has zero gross margin. Every dollar customers pay goes straight to Anthropic for Claude API calls and OpenAI for GPT-4 access. Then the company still needs to cover salaries, offices, servers, and other operating expenses.
An investor told tech analyst Ed Zitron that “Cursor is spending 100% of its revenue on Anthropic.” Zitron reported one user burned nearly $4 in AI costs just creating a simple to-do list, showing how unpredictable LLM consumption can be.
For context, Anthropic itself spent $2.66 billion on AWS against $2.55 billion in revenue through September 2025. Even the foundation model providers barely break even.
How Cursor Plans to Reach Profitability
The company has a plan to fix the profitability problem. It centers on building proprietary AI models that cost less to run than using Anthropic and OpenAI’s APIs.
In October 2025, Cursor launched Composer, their first proprietary coding model. According to the company, Composer runs 4 times faster than competing frontier models and works specifically for multi-file, multi-step coding tasks.
Composer trains on over 1 billion lines of code that Cursor users generate daily. As more developers use Cursor, the model gets better. As the model gets better, more developers want to use Cursor. This creates a data flywheel that competitors can’t easily replicate.
If proprietary models work, Cursor could cut AI costs from 100% of revenue down to 30-40%. Combined with higher prices and more enterprise customers paying premium rates, the path to profitability becomes clearer.
The company also launched Bugbot in July 2025, a debugging tool sold as a $40 per user monthly add-on. Additional products like this create new revenue streams beyond the core editor.
The Controversial Pricing Change
Cursor changed how it charges users in June 2025, creating significant backlash from developers.
Before June, the Pro plan was simple: $20 monthly got you 500 requests. Some advanced models like Sonnet cost 2 requests instead of 1, but spending was predictable. After hitting the limit, requests became slower but you never paid extra.
On June 16, 2025, Cursor switched to a credit system. The $20 monthly fee now includes a $20 credit pool. Different AI models consume different amounts of credits based on their actual API costs. Users can exceed their pool and get charged for overages.
The problems started immediately. Developers got surprise bills they didn’t expect. There was no live usage tracker showing credit consumption in real-time. Cursor didn’t communicate the change well. Many users thought “unlimited” applied to all models, not realizing it only covered Auto mode.
Some users reported burning through their entire monthly allocation after just 2-3 complex prompts using advanced models. Others complained that what was previously 500 requests became approximately 225 Claude requests under the new system.
Cursor CEO Michael Truell apologized on July 7, 2025: “We recognize that we didn’t handle this pricing rollout well and we’re sorry. Our communication was not clear enough and came as a surprise to many of you.”
Cursor responded by issuing refunds for affected users between June 16 and July 4, 2025, and publishing a clarification of the pricing. They added better usage dashboards. But trust with the developer community took a hit.
Despite the controversy, Cursor kept the credit system because it better aligns with their actual costs. Running premium AI models is expensive and variable. The credit system passes those costs to heavy users rather than subsidizing them through everyone’s subscriptions.
How Cursor Grew So Fast
Getting from zero to $1 billion ARR in 24 months required several things going right at once.
Building on VS Code was strategic. Developers already know VS Code, which has 70% market share among programmers. Cursor forked VS Code, meaning users get a familiar interface with zero learning curve. All VS Code extensions work in Cursor automatically.
The free tier converts exceptionally well. Cursor achieves a 36% freemium conversion rate, compared to the typical 2-5% for SaaS products. Free users get real value from unlimited basic autocomplete, then naturally hit upgrade triggers when they want premium features.
Word-of-mouth marketing worked. Cursor reached $100 million ARR without spending money on ads or building a sales team. All growth came organically from developers sharing the product on Twitter, Reddit, Hacker News, and YouTube.
Timing was perfect. ChatGPT launched in November 2022 and proved developers wanted AI help with coding. Cursor launched in 2023 as the first AI-native IDE just as this demand exploded.
The product actually works. Tab autocomplete is instant. Multi-file edits succeed reliably. The AI hallucinates less than alternatives. When developers try Cursor, they stick with it.
Cursor vs GitHub Copilot
GitHub Copilot is Cursor’s main competitor. Comparing the two shows why Cursor generates more revenue per user despite having fewer total users.
Cursor has about 2 million users with over 1 million paying. That generates $1 billion ARR, or roughly $500 average revenue per user annually.
GitHub Copilot has 20 million total users with 1.3 million paying subscribers as of July 2025. With individual plans at $10 monthly and business plans at $19 monthly, Copilot generates approximately $2 billion ARR according to Microsoft. That works out to about $1,538 per paying user annually.
However, Microsoft reported in 2024 that GitHub Copilot had become “a larger business than all of GitHub was” when Microsoft acquired it in 2018, suggesting significant revenue generation.
Why does Cursor monetize differently? Higher pricing is part of it. The Pro plan costs $20 versus Copilot’s $10, and Teams costs $40 versus Copilot’s $19.
But the bigger factor is positioning. Cursor’s credit system extracts more revenue from power users who consume lots of AI compute. Copilot’s flat pricing means Microsoft subsidizes heavy users with revenue from light users.
Also, Cursor is a complete IDE while Copilot is a plugin. Capturing the entire coding workflow lets Cursor monetize more of the value developers get from AI assistance.
Competition Is Heating Up
Despite Cursor’s strong position, competition is intensifying from well-funded players.
GitHub Copilot has Microsoft behind it, 1.3 million paying users, and distribution through GitHub. Microsoft could bundle Copilot free with VS Code subscriptions or GitHub accounts. The product is used by 90% of Fortune 100 companies.
Claude Code launched in May 2025 and already generates a $500 million run-rate according to Anthropic. It’s made by Anthropic, the same company providing Cursor’s AI models. Vertical integration could give Claude Code advantages.
Windsurf reached $82 million ARR by July 2025 with strong team collaboration features. It costs $30 monthly, undercutting Cursor on price. OpenAI acquired Windsurf for approximately $3 billion, validating the competitive threat.
Replit Agent is browser-based and focuses on generating complete applications rather than just editing code. It serves a different use case but competes for developer attention.
Cognition AI (Devin) raised $400 million at a $10.2 billion valuation in September 2025, focusing on autonomous AI software development.
The competitive threats are real. Microsoft could use its distribution advantage to overwhelm Cursor. Anthropic could make Claude Code the default choice. Foundation model prices might drop faster than Cursor can build proprietary alternatives.
What Analysts Project for Cursor Revenue
Based on current trajectory, analysts see Cursor’s revenue growing significantly but at a slowing pace.
For 2026, most estimates put Cursor revenue between $2 billion and $3 billion ARR. This assumes growth slows from the early-2025 pace of doubling every two months to a more sustainable 100-150% year-over-year growth rate.
By 2027, revenue could reach $4-5 billion as growth continues slowing to 50-80% annually. Competitive pressure increases and market penetration deepens.
By 2030, Cursor might generate $10-15 billion in annual revenue as a mature SaaS company growing 25-30% yearly. At that scale, profitability becomes more realistic.
The company could pursue an IPO in 2027-2028. Public market valuations for AI companies currently range from 12-15 times revenue for fast-growing businesses. That would put Cursor’s IPO valuation somewhere between $80 billion and $150 billion.
Three Possible Outcomes
Looking ahead, Cursor faces three main scenarios.
IPO has 40% probability. If Cursor reaches profitability or shows a clear path there, going public in 2027-2028 makes sense. The company would need to maintain growth while improving margins. An IPO at $80-150 billion valuation is realistic if fundamentals support it.
Acquisition has 30% probability. Microsoft, Google, Meta, or Salesforce might buy Cursor before an IPO. Price would likely be $40-60 billion. Microsoft makes the most sense given GitHub’s importance to their developer strategy. OpenAI reportedly approached Cursor about acquisition in 2024 but talks didn’t progress.
Staying private has 30% probability. With $3.3 billion raised, Cursor could keep building as a private company. They might raise more funding at $50-100 billion valuations and delay going public until 2030 or beyond.
Which outcome happens depends largely on whether proprietary models solve the profitability problem. If Composer works and cuts AI costs significantly, Cursor becomes more valuable and independent. If costs stay at 100% of revenue, acquisition becomes more likely.
The Risks to Watch
Every high-growth company faces risks. For Cursor, several factors could affect the $29 billion valuation.
AI technology changes fast. Features that differentiate Cursor today might become standard in foundation models tomorrow. The company needs to keep innovating ahead of the curve.
Balancing open-source popularity with commercial revenue is tricky. Cursor’s tools need to stay useful for free users while convincing companies to pay for premium features.
Competition from well-funded startups and big tech companies could pressure both market share and pricing. Microsoft has deep pockets and distribution advantages.
The AI agent coding market is still new. It might take longer than expected for companies to widely adopt AI-powered development tools, slowing Cursor’s enterprise growth.
Foundation model prices are dropping quickly. If Claude and GPT costs fall 50% in the next year, that hurts Cursor’s value proposition for building proprietary models.
What This Means for AI Developer Tools
The Cursor revenue story signals where AI tooling is headed. It shows that developers will pay significant money for AI that works reliably in their workflows.
It validates the “picks and shovels” approach to AI investing. Cursor doesn’t build foundation models. They build tools that make existing AI models useful for developers. This can be a better business because you benefit from all AI model improvements without betting on one specific approach.
For other AI startups, Cursor’s success demonstrates the importance of product-led growth. Letting developers use the product for free creates organic adoption that converts to paid enterprise contracts.
The $29 billion valuation also shows investors believe AI coding assistance is a massive market. If one tool can reach $1 billion revenue in 24 months, the total market size must be enormous. The AI coding assistant market was valued at $4.9 billion in 2024 and is expected to exceed $30 billion by 2032.
The Bottom Line on Cursor Revenue
Cursor generates $1 billion in annual revenue but loses $150 million because AI costs consume all revenue and more.
The company has a clear path forward. If proprietary models work, gross margins improve from 0% to 60-70%. Combined with higher prices and more enterprise customers, Cursor could be highly profitable by 2027-2030.
If proprietary models don’t work fast enough, acquisition becomes more likely. Microsoft or another tech giant might buy Cursor for $40-60 billion to own the category.
For now, Cursor has $3.3 billion in funding to execute the plan. The company has grown to over 300 employees and serves millions of developers across over 50,000 teams globally. The next two years are crucial. The company needs to prove that Composer reduces AI costs significantly while maintaining the product quality that drove the initial growth.
Whether the $29 billion valuation proves accurate depends entirely on whether Cursor can turn explosive growth into a sustainable, profitable business. The revenue is real and growing fast. The question is whether the economics can ever work at scale.