Is Cursor Profitable Today?

We research AI Wrappers every day, if you're building in this space, get our report

Cursor has become one of the hottest AI coding tools in Silicon Valley, raising nearly $1 billion at a jaw-dropping $9.9 billion valuation.

But beneath the explosive growth lies a fierce debate about whether the company actually makes money.

Industry analysts, investors, and users are openly questioning if Cursor's economics work at all.

Here's what the data and reports actually show.

In our 200+-page report on AI wrappers, we'll show you the real user pain points that don't yet have good solutions, so you can build what people want.

What revenue numbers has Cursor officially disclosed?

Cursor hit $500 million in annualized recurring revenue by June 2025.

This made Anysphere the fastest-growing AI company outside of OpenAI and Anthropic. The company reached $100 million ARR in just 12 months and doubled that to $200 million by March 2025. Bloomberg sources confirmed the $500 million figure represented a 60% jump from the $300 million reported in mid-April 2025.

Cursor charges users $20 per month for Pro plans and $40 per month for business subscriptions.

By early 2025, the company had over 360,000 paying subscribers and serves more than 1 million daily active users. The company's growth trajectory shows ARR doubling approximately every two months during peak periods, making it one of the most rapid revenue expansions in SaaS history.

These figures come from Bloomberg, TechCrunch, and company sources speaking to trade publications.

Does Cursor claim to be profitable?

Cursor has never publicly claimed profitability.

CEO Michael Truell has focused interviews on product vision and growth metrics rather than financial sustainability. One analysis suggests "profitability expected in late 2026" based on the company's trajectory, but this remains speculative and unconfirmed by official statements.

The company raised $900 million at a $9.9 billion valuation in June 2025.

This massive funding round suggests investors believe in long-term potential despite current economics. Anysphere operates with around 150 employees and reportedly maintains single-digit monthly cash burn, which sounds positive but doesn't account for the full cost structure including compute expenses paid to AI providers.

The absence of profitability claims from a company generating $500 million ARR is itself telling in an industry where founders typically celebrate any positive margin milestones.

What do Cursor users say about profitability on Reddit and forums?

Reddit discussions show deep skepticism about whether Cursor can survive with its current economics.

Multiple threads questioned if Cursor is "actually making 100% profit" or if it's burning through investor cash. Users pointed out the contradiction between explosive growth and reports of 100% revenue going to Anthropic. Developer forums filled with calculations showing individual subscriptions can't possibly cover the compute costs heavy users consume.

Some users defended Cursor by noting enterprise customers likely have better margins through usage limits and seat-based pricing.

Others argued the pricing crisis proves the fundamental business model is broken. One popular sentiment: if Cursor needs to charge users the true cost of compute, the product becomes too expensive for most developers to justify. The debate centers on whether Cursor can transition enough users to enterprise plans before investor patience runs out.

Power users who loved the original unlimited model feel betrayed and are actively exploring alternatives like Windsurf and Claude Code direct.

In our 200+-page report on AI wrappers, we'll show you which areas are already overcrowded, so you don't waste time or effort.

Is Cursor profitable according to industry analysts?

What are analysts saying about Cursor's margins?

No major analyst reports suggest Cursor is currently profitable.

Multiple sources confirm Cursor operates with negative gross margins, meaning it loses money on every sale before accounting for operating expenses.

Ed Zitron concluded that Cursor is "deeply unprofitable" and was so even before Anthropic raised prices.

Mayfield VC noted that companies relying heavily on third-party models "often scale costs faster than revenue" and can't hit the 80%+ gross margins you see in normal SaaS businesses.

What did Tom Dotan reveal about Cursor's costs?

Tom Dotan of Newcomer reported in August 2025 that an investor told him "Cursor is spending 100% of its revenue on Anthropic."

This means every dollar Cursor earns immediately goes to pay for AI models, leaving nothing for employees or infrastructure.

Multiple sources familiar with the financials confirmed to Dotan that Cursor has negative gross margins, with different customer types having "wildly different cost profiles" that bounce around week by week.

Investment firm Foundamental ran the numbers and found Cursor pays approximately $650 million annually to Anthropic while generating roughly $500 million in revenue, creating a negative 30% gross margin.

How much does Cursor spend on Amazon Web Services?

Ed Zitron obtained exclusive AWS billing data showing Cursor's cloud costs doubled from $6.2 million in May to $12.6 million in June 2025.

These figures represent only the portion of Anthropic services purchased through AWS, not total compute spending. The timing matched exactly when Anthropic launched Priority Service Tiers in late May, forcing Cursor into expensive guaranteed throughput contracts.

Zitron confirmed all numbers represent final cash paid after discounts and credits, verified with multiple sources.

For context, Anthropic itself spent $2.66 billion on AWS against an estimated $2.55 billion in revenue through September 2025.

What does Ed Zitron's investigative work show?

Zitron's reporting uses Cursor as his main example showing how AI application companies are "antithetical to the way that software is sold."

He documented a Cursor user burning nearly $4 in minutes creating a simple Todo file, showing how LLMs burn tokens unpredictably.

Zitron believes "generative AI is, at its core, unprofitable" and that no company building on Anthropic or OpenAI "has a path to profitability outside of massive, unrealistic price increases."

His analysis suggests the entire AI application layer just moves money from VCs to cloud providers, with only Amazon, Google, and Microsoft winning.

What feedback are users sharing publicly?

Derick David's Medium article documented user complaints that Cursor "has fallen behind" competitors despite its $9.9 billion valuation.

Published in August 2025, the piece captured viral Reddit discussions where developers systematically compared Cursor against alternatives. ClickUp's testing found Cursor "tends to lag when handling complex, multi-file projects, making it inefficient for larger development teams."

Multiple independent reviews documented that Cursor "sometimes generates irrelevant or incorrect code snippets, which disrupts development flow."

The article suggested pricing pressures forced Cursor to cut quality to manage costs.

What's the bottom line from analysts?

Every new customer who actively uses the product increases infrastructure costs proportionally, unlike traditional software where adding users costs almost nothing.

Analyst Chris Paik wrote that Cursor faces a "double whammy." Users expect frontier model performance, which only Anthropic and OpenAI provide, while those same providers are also Cursor's fiercest competitors.

The problem stems from power users who consume far more compute than their subscription fees cover.

Individual developers on $20 or $40 monthly plans can rack up hundreds in costs that Cursor absorbs.

The only path to profitability analysts see is aggressive enterprise adoption with consumption-based pricing that passes costs directly to customers.

In our 200+-page report on AI wrappers, we'll show you which ones are standing out and what strategies they implemented to be that successful, so you can replicate some of them.

How does Cursor's relationship with Anthropic affect profitability?

Cursor is Anthropic's largest customer, creating a dependency that makes profitability nearly impossible under current terms.

Anthropic earns $1.4 billion of its annualized revenue from its top two customers, with Cursor being number one. This gives Anthropic huge pricing power. When they introduced Priority Service Tiers in May 2025, Cursor had no choice but to accept dramatically higher costs or risk service degradation.

Making matters worse, Anthropic uses revenue from Cursor to fund Claude Code, a direct competitor to Cursor's core product.

The relationship is what Ed Zitron calls the "Subprime AI Crisis." Model providers offer services at a loss, then raise prices on customers who are also unprofitable.

Cursor hired two leaders from Anthropic's Claude Code team in 2025, including former lead developer Boris Cherny. This suggests Cursor sees the existential threat and is trying to reduce dependency through proprietary models. But developing competitive in-house models requires hundreds of millions in R&D that Cursor can't afford while losing money on every transaction.

Anthropic charges $15 per million input tokens and $75 per million output tokens for Claude Opus 4, far higher than competitors.

Can Cursor become profitable with enterprise customers?

Enterprise pricing offers Cursor's only realistic path to profitability, but the transition is far from guaranteed.

Sources told Tom Dotan that "enterprise pricing model has far better margins because it's a combination of seat based and consumption based pricing."

Over 14,000 businesses were paying customers as of mid-2025.

Enterprise customers accept usage limits and per-seat fees that individual developers won't tolerate, letting Cursor avoid subsidizing power users who consume thousands in compute monthly.

However, enterprise sales require building teams that Cursor has barely started developing.

The company historically focused on viral bottom-up adoption through individual developers. Shifting to enterprise requires dedicated sales, implementation support, and compliance capabilities. All expensive before generating returns. Moreover, Microsoft bundles GitHub Copilot into enterprise agreements at effectively zero marginal cost, giving Cursor a competitor with way better economics.

Even if enterprise saves Cursor, it means abandoning the individual developer market that built the brand.

In our 200+-page report on AI wrappers, we'll show you the real challenges upfront - the things that trip up most founders and drain their time, money, or motivation. We think it will be better than learning these painful lessons yourself.

Why do investors value Cursor at $9.9 billion if it's unprofitable?

Investors are betting on Cursor capturing the massive developer tools market even if current economics are broken.

The $9.9 billion valuation from June 2025 reflects belief that Cursor can transition to profitable enterprise customers and reduce dependency on expensive third-party models. Thrive Capital, Andreessen Horowitz, and other top VCs see Cursor as the category leader in AI coding—a market potentially worth tens of billions. The fastest revenue growth in SaaS history gives investors confidence that user demand is real.

Venture capital in 2025 operates with excess capital needing deployment and tolerance for negative unit economics during "land grab" phases.

The frothy market conditions mean investors chase growth over profitability, hoping for eventual market consolidation. Negative gross margins become acceptable if you believe Cursor can achieve monopoly pricing power. The scooter craze of 2017-2020 showed billions flowing into fundamentally broken economics, and AI might be following the same pattern.

Some investors may also be calculating that acquisition by Microsoft, Google, or another tech giant offers an exit before profitability becomes necessary.

Will Cursor's costs decrease as AI models improve?

No evidence suggests compute costs will decrease significantly despite AI boosters claiming "costs will come down."

Ed Zitron directly addresses this argument: "Please don't waste your breath saying 'costs will come down.' They haven't been, and they're not going to." Reasoning models that produce better outputs require more tokens, not fewer. The cost of inference is actually increasing because getting quality results demands longer context windows and more computational passes.

Anthropic and OpenAI are both losing billions annually despite their scale.

When Anthropic launched Claude Opus 4 and Sonnet 4 in May 2025, prices went up rather than down—Opus 4 costs $15 per million input tokens versus $3 for earlier models. Each new generation of frontier models costs more to run because the capabilities require more compute. The model providers charge more as their models get better, not less.

Even if chip efficiency improves, the demand for more capable models will eat up any savings rather than lowering prices for end users.

Why did Cursor change its pricing in June 2025?

Cursor switched from unlimited usage to consumption-based limits because newer AI models became too expensive to offer at flat rates.

On June 16, 2025, the company replaced its popular $20 Pro plan that offered 500 fast requests with a credit-based system providing $20 worth of API usage. Users who thought they had "unlimited" access suddenly faced hard limits and unexpected overage charges. The change followed Anthropic's May 22 launch of Claude Opus 4 and Sonnet 4 models, which consume significantly more tokens per request.

CEO Michael Truell admitted in a blog post that "we didn't handle this pricing rollout well, and we're sorry."

The company offered full refunds for unexpected charges between June 16 and July 4. But the damage was done—Reddit and developer forums exploded with complaints about hitting limits after just a few prompts. One user reported burning $2 in credits with only two agent requests, while another faced $71 in charges for a single day of coding.

Cursor simultaneously launched a $200 monthly Ultra plan with 20x more usage, making the transition feel like a bait-and-switch to longtime users.

Is Cursor representative of broader AI startup problems?

Yes, Cursor shows the same profitability challenges facing the entire AI application layer.

Ed Zitron notes "my gut instinct is that this is an industry-wide problem." Perplexity spent 164% of its revenue in 2024 on AWS, Anthropic, and OpenAI combined. OpenAI itself spent 50% of revenue on inference and 75% on training, ending up spending $9 billion to lose $5 billion in 2024. Every single company offering generative AI services outside training data providers is reportedly losing money.

The pattern shows AI startups are basically just moving venture capital to cloud providers without retaining value.

Traditional software enjoys near-zero marginal costs, but AI flips this model. Each new user costs real money. Mayfield's research shows 92% of AI software companies now use mixed pricing models trying to manage unit economics. The shift from unlimited to consumption-based pricing across the industry signals widespread acknowledgment that the original models were unsustainable.

Cursor isn't an outlier but rather the most visible example of a problem infecting hundreds of AI startups simultaneously.

In our 200+-page report on AI wrappers, we'll show you what successful wrappers implemented to lock in users. Small tweaks that (we think) make a massive difference in retention numbers.