The AI Coding Darling's Financials Tell a Different Story

Cursor has become the poster child for AI-native development tools. The company reached $1 billion in annual recurring revenue faster than any SaaS company in history—just 24 months. Its valuation rocketed from $400 million in August 2024 to $2.6 billion in December 2024 to $9.9 billion in June 2025 to a staggering $29.3 billion in November 2025. That's a 1,026% valuation increase in one year.

But according to Market Clarity's investigation, there's a devastating problem hiding behind the growth: Cursor is paying more to Anthropic for Claude API access than it's generating in revenue. The company has negative gross margins.

The Anthropic Dependency Problem

According to Sacra's financial analysis, the numbers are stark:

• $650 Million Annual Anthropic Spend: Investment firm Foundamental estimates Cursor pays approximately $650 million annually to Anthropic for Claude API access.

• $500 Million Annual Revenue: Meanwhile, Cursor generates roughly $500 million in revenue.

• Negative 30% Gross Margin: That's right—for every dollar Cursor makes, they spend $1.30 on just their primary AI provider. This doesn't include salaries, infrastructure, marketing, or any other costs.

• "100% of Revenue on Anthropic": Tom Dotan of Newcomer reported in August 2025 that an investor told him "Cursor is spending 100% of its revenue on Anthropic." Multiple sources confirmed this to Dotan.

"Cursor is spending 100% of its revenue on Anthropic. Multiple sources confirmed that Cursor has negative gross margins. This isn't a path to profitability—it's a path to becoming an Anthropic subsidiary." — Market Clarity Analysis

The July 2025 Pricing Debacle

When your costs exceed your revenue, you have two options: cut costs (difficult when your product IS the API calls) or raise prices. Cursor chose the latter, and according to Wikipedia's documentation, it backfired spectacularly:

• Request Caps to Credits: In July 2025, Cursor switched its $20 Pro plan from 500 requests to a usage-metered credit pool system.

• Unexpected Charges: Users reported unexpected charges as their workflows consumed more credits than the old request system.

• Public Backlash: The complaints were loud enough that Cursor was forced to roll back limits and promise refunds.

• Auto Mode Downgrade: The new system forces users who exhaust their credit pool to switch to "Auto mode" (lower quality) or pay overage at API cost.

The "Sam" Customer Service Disaster

In April 2025, according to AI Certs' reporting, Cursor's AI help-desk program "Sam" caused another controversy:

• Invented Policy: The AI customer service bot invented a non-existent login policy and communicated it to users as company policy.

• User Cancellations: Users cancelled subscriptions based on the fabricated policy.

• Staff Intervention: Human staff had to intervene and issue refunds.

• Trust Damage: The incident raised questions about Cursor's operational maturity despite its valuation.

The Valuation Reality Check

According to Digital Applied's analysis, Cursor's $29.3 billion valuation prices in perfection:

• 60x Revenue Multiple: At $500M revenue and $29.3B valuation, Cursor trades at roughly 60x revenue—comparable to the most optimistic SaaS valuations of 2021.

• No Path to Profitability Disclosed: The company hasn't publicly explained how it will achieve positive margins given its cost structure.

• Preferred Share Terms: Private valuations rely on preferred share terms that may not translate at exit. Paper wealth doesn't equal realized wealth.

• "Profitability Expected Late 2026": One analysis suggests profitability in late 2026, but this is speculative and depends on either dramatic cost reductions or price increases.

"This round prices in perfection with minimal room for execution error. Any slowdown in seat expansion, enterprise budget consolidation in 2026, or successful competitive response from Microsoft's deeper-pocketed GitHub Copilot would compress multiples rapidly." — Industry Analysis

The Competition Problem

Cursor doesn't exist in a vacuum. According to SaaStr's analysis, competition is intensifying:

• GitHub Copilot: Microsoft's offering has deeper pockets, enterprise relationships, and GitHub integration advantages.

• Claude Code: Anthropic's own CLI tool threatens to disintermediate Cursor by going direct to developers.

• Windsurf: Codeium's AI IDE is gaining traction as an alternative.

• Open Source Options: Continue, Aider, and other open-source tools offer similar functionality without subscription costs.

User Complaints: "Cursor Has Fallen Behind"

According to CTOL Digital's coverage, user sentiment has shifted:

• Viral Reddit Discussions: A Medium article by Derick David documented widespread Reddit complaints that Cursor "has fallen behind" competitors despite its valuation.

• Quality Concerns: Users report declining output quality as the company scales.

• Support Issues: The "Sam" incident wasn't isolated—support quality has become a common complaint.

• Price Sensitivity: After the July pricing changes, users are more conscious of value-for-money.

What Happens If Anthropic Raises Prices?

Cursor's business model has an existential vulnerability: Anthropic controls its cost structure. If Claude API prices increase—or if Anthropic decides to compete more directly with Claude Code—Cursor's economics get even worse.

• No Alternative: Cursor is built on Claude. Switching to GPT or another model would require significant engineering work and might degrade product quality.

• Negotiating Leverage: Cursor is one of Anthropic's largest customers, which provides some protection. But Anthropic's own interests may not align with Cursor's survival.

• Integration Risk: If Anthropic's Claude Code becomes good enough, why would developers pay for a middleman?

The Bull Case (Such As It Is)

To be fair, Cursor bulls argue:

• First-Mover Advantage: Cursor has mindshare and habits that are hard to break.

• Product Quality: When it works, Cursor's integration is genuinely excellent.

• Enterprise Momentum: Large organizations are adopting Cursor, and enterprise customers are sticky.

• Cost Improvements Coming: AI inference costs are dropping rapidly; margins should improve over time.

The Bottom Line: A Great Product, A Questionable Business

Cursor is a genuinely good product. Developers who use it report real productivity improvements. The experience of AI-assisted coding in Cursor can feel magical.

But a good product isn't the same as a good business. At $29.3 billion, Cursor's valuation implies it will become one of the most valuable software companies ever created. That requires achieving positive margins, defending against well-funded competitors, and maintaining its relationship with a supplier who could become a competitor at any time.

For developers, keep using Cursor if it works for you—but don't build workflows that can't survive if the company struggles. For investors, the valuation leaves no room for error in what is clearly an error-prone business model. And for the industry, Cursor is a case study in how fast growth can hide fundamental business problems until they become impossible to solve.