$2B for what’s basically an AI coding assistant is kind of wild.
Feels like we’ve officially entered the era where dev tools are becoming full-blown platforms instead of just utilities.
Curious how much of that revenue is hype vs long-term retention, though subscriptions are easy to start, harder to justify after the novelty wears off.
Ciappatos on
In reality terms this means they had one 30 day period where they made 166m
Hormones-Go-Hard on
This is crazy because cursor is so far behind codex and Claude code. Who still uses cursor??
Kyouhen on
Friendly reminder that annualized revenue is taking your best month and multiplying it by 12. By definition they made less than $2 billion, they just aren’t willing to tell us how much less. And with the way the AI space works odds are this is all just them signing a deal with Nvidia or something. That’s like me selling my car and declaring that I made $360k this year.
InternationalToe3371 on
$2B annualized is insane for a dev tool tbh.
Shows how big the “AI for builders” wave really is. If you sit inside the workflow, you win.
Also kind of proof that distribution > pure model quality.
Curious how durable that revenue is once competition catches up though.
neuenono on
And what’s their annualized burn?
LargeSinkholesInNYC on
Cursor is a shit company.
ShadowBannedAugustus on
Ok so to the devs that will be here:
I use copilot chat in agent mode within vs code, usually with Sonnet 4.6 as my model of choice. I like it a lot. How far behind am I?
tmdblya on
Never heard of it
zmvf on
Claude code and codex shows that companies like Cursor don’t have a future. Why use a middleman when the products directly from AI labs are better
textonic on
I don’t get the hate. It allows you model flexibility, uses less tokens compared to Claude to and for most people that’s great bonus’
startfragment on
With no real moat dug yet.
phejster on
Uh huh, ok sure
pfc-anon on
They have no moat or edge. It’s a vscode fork which was already very bloated. And they don’t share what their burn rate is, they are routing calls to all the frontier models, I’m sure they’re taking a substantial haircut on that alone. To add to that, if their goal is to replace devs eventually then their business model is flawed.
In other words, ripe for acquisition, someone will acquire them for the developer eyeballs eventually.
Kbartman on
Wonder how much revenue the apps which used it made
_ECMO_ on
So a service that apparently nearly every coder uses has only negligible revenue. Wow.
AtraVenator on
An LLM and VS code wrapper. Folks are mad.
Claude Code / Codex cli and VSCode does the same thing for you.
immutate on
The company I work at is throwing money at Cursor, without anyone realizing that vscode has all the same capabilities, and fewer issues. Cursor provides better metrics for them to track usage, though, which is what they really want.
jawndell on
Doesn’t Amazon, Google, et al. have similar tools in house?
CanvasFanatic on
Now do their annual expenditures.
Leave A Reply
Du musst angemeldet sein, um einen Kommentar abzugeben.
20 Kommentare
$2B for what’s basically an AI coding assistant is kind of wild.
Feels like we’ve officially entered the era where dev tools are becoming full-blown platforms instead of just utilities.
Curious how much of that revenue is hype vs long-term retention, though subscriptions are easy to start, harder to justify after the novelty wears off.
In reality terms this means they had one 30 day period where they made 166m
This is crazy because cursor is so far behind codex and Claude code. Who still uses cursor??
Friendly reminder that annualized revenue is taking your best month and multiplying it by 12. By definition they made less than $2 billion, they just aren’t willing to tell us how much less. And with the way the AI space works odds are this is all just them signing a deal with Nvidia or something. That’s like me selling my car and declaring that I made $360k this year.
$2B annualized is insane for a dev tool tbh.
Shows how big the “AI for builders” wave really is. If you sit inside the workflow, you win.
Also kind of proof that distribution > pure model quality.
Curious how durable that revenue is once competition catches up though.
And what’s their annualized burn?
Cursor is a shit company.
Ok so to the devs that will be here:
I use copilot chat in agent mode within vs code, usually with Sonnet 4.6 as my model of choice. I like it a lot. How far behind am I?
Never heard of it
Claude code and codex shows that companies like Cursor don’t have a future. Why use a middleman when the products directly from AI labs are better
I don’t get the hate. It allows you model flexibility, uses less tokens compared to Claude to and for most people that’s great bonus’
With no real moat dug yet.
Uh huh, ok sure
They have no moat or edge. It’s a vscode fork which was already very bloated. And they don’t share what their burn rate is, they are routing calls to all the frontier models, I’m sure they’re taking a substantial haircut on that alone. To add to that, if their goal is to replace devs eventually then their business model is flawed.
In other words, ripe for acquisition, someone will acquire them for the developer eyeballs eventually.
Wonder how much revenue the apps which used it made
So a service that apparently nearly every coder uses has only negligible revenue. Wow.
An LLM and VS code wrapper. Folks are mad.
Claude Code / Codex cli and VSCode does the same thing for you.
The company I work at is throwing money at Cursor, without anyone realizing that vscode has all the same capabilities, and fewer issues. Cursor provides better metrics for them to track usage, though, which is what they really want.
Doesn’t Amazon, Google, et al. have similar tools in house?
Now do their annual expenditures.