Cursor, the AI code editor that launched three years ago, is in talks to raise $2 billion at a valuation north of $50 billion. The round does not even include the new capital. Post-money, you are looking at $52 billion for a company that makes a text editor.

Andreessen Horowitz and Thrive Capital are co-leading. Nvidia is a strategic co-investor. The round is already oversubscribed. Five months ago, the company was worth $29.3 billion. It has nearly doubled since November.

For context: Ford is worth $37 billion. Delta Air Lines is worth $30 billion. Marriott International is worth $44 billion. Cursor, which employs a fraction of their workforce and has existed for roughly 1,100 days, is worth more than any of them.

The Numbers That Got Everyone's Attention

Cursor crossed $2 billion in annual recurring revenue. That is $2B ARR in three years from a standing start. It raised $900 million in June 2025, then $2.3 billion in November 2025, and now $2 billion more. That is over $5 billion raised in under 12 months for a company that primarily sells software subscriptions to developers.

The valuation math says investors are pricing Cursor at roughly 25x revenue. That is aggressive even by AI standards. But the pitch is simple: AI coding is the first AI use case with undeniable, measurable value. Developers who use Cursor ship faster. Companies that ship faster win. The entire global software development market is the TAM.

The Problem: Everyone Wants This Market

Cursor was the first mover in AI coding agents, but the field is now a bloodbath. Anthropic rebuilt Claude Code into a full desktop IDE with persistent routines. OpenAI launched Codex as a standalone product inside ChatGPT Pro at $100 per month. Google is weaving Gemini into every developer tool it touches. GitHub Copilot, backed by Microsoft and OpenAI, already has tens of millions of users.

The question at $50 billion is not whether Cursor is good. It is whether a standalone coding tool can defend its position against companies that control the underlying models. Anthropic makes Claude. OpenAI makes GPT. Google makes Gemini. Cursor rents from all of them. That is a distribution advantage today and a vulnerability tomorrow.

Nvidia's participation is the most telling signal. Jensen Huang is not investing in a text editor. He is investing in the application layer that drives compute demand. Every Cursor query burns GPU cycles. More Cursor users means more chips sold. Nvidia does not care who wins the coding war as long as someone keeps buying hardware.

What to Watch

AI coding is the first sector where AI has a clear, undeniable, revenue-generating use case. We covered this two weeks ago when xAI started renting tens of thousands of GPUs to Cursor because its own utilization was "embarrassingly low." Cursor is now one of the single largest GPU consumers in the entire AI ecosystem, and it does not even build models.

The $50 billion bet is that AI coding becomes as essential to software development as the internet became to communication. If that is true, this valuation will look cheap in two years. If the model providers decide to cut out the middleman, it will look like the most expensive text editor ever built.

First reported by CNBC. Bloomberg first reported the fundraise was in progress.