Cursor Is Worth $50 Billion Now. The Revenue Chart Explains Why, and the Worry.
Anysphere's Cursor is raising about $2B at a $50B valuation, double its November mark. Revenue went from $100M to $2B in roughly a year. The growth is real. So is the question of what it's defending.
By Priya Raman · · 3 min read
query://industryStart with the one chart that explains the whole story. Cursor's annualized revenue went from $100 million in January 2025 to $500 million by June, $1 billion by November, and $2 billion by February 2026. The company forecasts ending this year above a $6 billion run rate.
That curve is why Anysphere, the company behind the Cursor code editor, is now raising around $2 billion at a roughly $50 billion valuation, co-led by Andreessen Horowitz and Thrive Capital with Nvidia as a strategic investor. The round is oversubscribed, and it nearly doubles the $29.3 billion valuation Cursor reached just five months earlier. When revenue doubles, the valuation doubling is not the strange part.
What the number is actually buying
A $50 billion valuation on a company built by four MIT graduates in three years is the kind of thing that sounds absurd until you look at the revenue, and then sounds merely aggressive. At a $6 billion forward run rate, the multiple is high but not unhinged by current AI standards. The market is paying for the growth rate, and the growth rate has, so far, delivered.
Here is what gives me pause, and it has nothing to do with the multiple.
Cursor sells an AI coding experience built on top of models it does not own. The intelligence comes from frontier labs, the same labs now shipping their own coding tools. Anthropic has Claude Code. Google just pushed everyone toward its Antigravity platform. Microsoft owns GitHub Copilot and its new metered pricing. Every one of Cursor's suppliers is also a competitor, and several of them can afford to lose money on coding tools indefinitely to win the developer relationship.
The moat question
Cursor's defenders will tell you the moat is the product: the editor experience, the way it handles large codebases, the taste in a thousand small decisions that make developers stay. That is real, and I do not want to wave it away, because product quality has kept Cursor ahead of better-funded rivals for two years already.
But "we are better at the experience" is a moat that has to be re-earned every model generation, against opponents who supply your core ingredient and have unlimited budgets. The bet a $50 billion valuation makes is that Cursor can keep out-executing the companies it depends on, forever. Sometimes that bet pays off. Sometimes the platform owner decides the application layer is too valuable to leave to a customer.
Where it could crack
The revenue trajectory is the bull case, and it is genuinely impressive, which is why the risks worth taking seriously are the ones that could break it rather than the valuation math. The thinnest spot is gross margin. Reselling another company's model intelligence at scale is a slimmer business than the top-line growth makes it look, and the token economics are brutal. Underneath that sits retention: two billion dollars of revenue is worth nothing if it churns to a free Microsoft bundle the next time a CFO trims the software budget. And all of it rests on the goodwill of suppliers who are also competitors. The day a frontier lab decides to undercut Cursor on price, the model gets its real stress test.
Cursor earned this valuation with the best revenue chart in software. Whether it keeps it depends on a question the chart cannot answer: what happens when the companies you are built on want your seat at the table.
enjoyed this one?_
Get the next one in your inbox.
One email every morning. The AI news that matters, decoded in five minutes.
up_next → Industry
Anthropic Just Passed OpenAI in Business Adoption. One Survey Says Don't Celebrate Yet.
The Ramp AI Index shows more US businesses now pay for Claude than ChatGPT for the first time. A separate IDC survey tells a more cautious story. Both are right, because they measure different things.
