I'm skeptical about this VSCode fork commanding a $3 billion valuation when it depends on API services it doesn't own. What's their moat here?
For comparison, JetBrains generates over $400 million in annual revenue and is valued around $7 billion. They've built proprietary technology and deep expertise in that market over decades.
If AI (terminology aside) replaces many professional software engineers and programmers like some of its fierce advocates say it would, wouldn't their potential customer base shrink?
Professionals typically drive enterprise revenue, while hobbyists—who might become the primary users—generally don't support the same business model or spending levels.
Part of what you're missing is that OpenAI needs to justify its own overinflated valuation. They raise money on the premise that an AI-native company can and will outcompete giant established players, so lowballing Windsurf would run counter to the narrative they're selling to their own investors.
The article also doesn't say that it's $3B in cash that OpenAI is spending. They might be giving Windsurf $3B worth of OpenAI shares - paying an inflated value for Windsurf with their own inflated value.
OpenAI just had a fundraising round that put them at $300B. Maybe they're just giving Windsurf 1% of OpenAI. Maybe they're even giving less than 1% - if OpenAI was worth $300B at the end of March and $150B last October, maybe they're worth $400B now. Maybe Windsurf is getting 0.75% of OpenAI that's "valued" at $3B.
> OpenAI just had a fundraising round that put them at $300B. Maybe they're just giving Windsurf 1% of OpenAI
That is the most hilarious maths I have ever seen, if this is true then it's maybe the biggest "holy fuck it's a bubble all the way down" I have ever seen
They built all of this assuming VSCode was a solid foundation for the next 30 years and I've completely undermined VSCode's technical foundations. Their castle is gonna sink into the swamp...
$400M in real revenue versus $300M in annual recurring revenue (ARR) are totally different things.
Real revenue is money actually earned, while ARR just multiplies one month's subscription revenue ($25M) by 12, ignoring customer churn.
Startups love flashing ARR figures because "$300M ARR" sounds impressive, but without knowing churn rates, they might never actually collect that full amount.
I don't think that distinction changes how much each of them is worth relatively to each other.
Achieving $300M ARR in 1 year is extremely extremely impressive regardless of churn or any other metrics really (assuming reasonable numbers). Being valued at $9B because of it doesn't seem out of line.
I'm skeptical of Cursor and not using Cursor myself. I actually use IntelliJ because I write Java.
Cursor's valuation is not unreasonable. But somehow you phrase it like $9B valuation for the fastest growing company that achieves the highest revenue per employee in the history of modern civilization is out of whack somehow.
Unless you have reason to believe the revenue is declining in recent months or will decline in near future, ARR is a better metric. last year real revenue made sense only for low growth companies.
Cursor just lost access to the extension marketplace and key proprietary plugins that they were using against Microsoft's terms, Windsurf has been eating a chunk of their mindshare, and Copilot is catching up.
That's three good reasons to believe that lots of people will be cancelling in the next months unless something changes.
The entire reason OpenAI has a high valuation is the expectation that AI will get a lot better in the next few years. If that happens, building a clone of Cursor/windsurf should be trivial. The only reason you would buy windsurf today is to either pump up the bubble OR use it to increase your market share of developers by taking users away from claude
I feel jetbrains is squandering an opportunity here. Cursor is significantly easier to build then any IDE in the jetbrains ecosystem. The technology jetbrains is very hard to replicate. While the technology cursor uses should be trivial to replicate.
If jetbrains can combine there IDE technology with cursor technology, that would be ideal.
I think the problem is jetbrains tech is sort of already very biased in a certain direction and it's hard for them to pivot as fast into this new AI direction.
JetBrains launched their cursor competitor a few weeks ago.
I prefer Claude Code still because it has access to more tools - Junie seems unable to fetch URLs and do other things. But that's a tiny gap that JetBrains can close quickly, and the Junie UI is quite pretty. Plus, inside the IDE they can equip the model with far more advanced tools than Claude Code will have from the CLI: inside Code Claude has to explore the codebase by banging stones together with ripgrep, whereas in the IDE it can be equipped with tools to access the indexes and navigate around like a human would.
In theory, JetBrains should be able to compete very strongly in this market. Their single line completion model is already excellent.
> If jetbrains can combine there IDE technology with cursor technology, that would be ideal.
Just give them some time, they're not stupid. I'd drop Cursor in an instant once JetBrains catches up, because IntelliJ IDEs are just a way more powerful.
They're giving out 1 month free if you're paying for their IDEs already. I've tried it last year and it was very limited, not "agentic". Now they've launched an agentic version called Junie and also gave another 1 month free, and I've tried it again.
It's a nice improvement over the last edition, but still quite not "smart" as Cursor or Windsurf. The agent seems too shortsighted compared to competitors: it may stop looking for files or making edits sooner and you're left with code made with incomplete context (that does not work or just doesn't address your needs). It also does not fix linter/compiler errors from its own output code before finishing, unlike Cursor.
Just consider what it fundamentally is: a company at the leading edge of a product category that has found absurdly strong technology/use-case fit, and is growing insanely fast.
Looking for a moat in the technology is always a bit of a trap - it’s in the traction, the brand awareness, the user data etc.
> Looking for a moat in the technology is always a bit of a trap - it’s in the traction, the brand awareness, the user data etc.
Traction, brand awareness, and user data do not favor Windsurf over GitHub Copilot. The few of us who follow all the new developments are aware that Windsurf has been roughly leading the pack in terms of capabilities, but do not underestimate the power of being bundled into both VS Code and GitHub by default. Everyone else is an upstart by comparison and needs some form of edge to make up for it, and without a moat it will be very hard for them to maintain their edge long enough to beat GitHub's dominance.
Definitely take that point. But this valuation is perhaps more about how much that traction, brand and data is worth to OpenAI, who cannot buy Copilot. $3bn doesn’t seem so disproportionate in that context especially given the amount of money being attracted to the space.
Define losing? My company pays for Copilot but not for Cursor, and it's not at all clear to me that we're the exception rather than the norm. What numbers and data are you working with?
That's not actually how unseating an incumbent works. The incumbent can adapt to the threat for quite a while if they act on it, they just have to not be Blockbuster. Copilot is showing every sign of making up ground feature-wise, which is bad news for the runners up.
Incumbent advantage of being in VS Code already? Thing is, Cursor is basically just VS Code, there's hardly any barrier to switching, so it's quite a weak advantage.
In brand velocity maybe, but copilot is rapidly reaching feature parity with cursor and will invariably overtake it—while costing less to users.
Same with Google vs OpenAI. I tend to agree with the sentiment that I most frequently hear which is that OpenAI is the currently popular brand, but that can only carry them so far against what will eventually be a better offering for cheaper.
Yeah it seems like there's really no "adult supervision" at all in OpenAI. This purchase was a panic move. Windsurf would be worthless without the AI. Probably OpenAI knows that AI is now a commodity technology and no longer a space they can monopolize so they're just trying to get off a ship that's sinking, and find some viable path to having a tech that doesn't ultimately depend on OpenAI even having a monopoly any longer.
If that's so, then why is Codex such an inferior product to Claude Code? And why haven't they already built an code editor or at least VS Code extension yet?
JetBrains has been making IDE for a decade. They were the only company that actually made money by selling IDE. So I assume they have the best programmers who understand IDE.
However they fail to make a Cursor competitor so far. This alone suggests it's a harder task than meets the eye.
It is, but you are assuming that only a well known IDE team would do it. To me JetBrains is the least likely to be an innovator here because they depend on their reputation for being a mature technology.
Someone like me isn't known at all but it means I have been able to experiment for a long time without pressure, which is how you do real innovation.
JetBrains as a company probably owns 10 million lines of code and it's just really hard to move fast when you're tugging that kind of ball and chain
If OpenAI just provides AI, then the various IDEs development wrappers / IDEs / low-code etc. can collectively bargain against OpenAI for low rates. If OpenAI has an alternative, then they can charge higher rates for all plugins/ etc. and give the market an alternative.
Enterprises won't require less software. If they require fewer software engineers, that would be those few engineers producing so much more software with better tools, for example, AI wrappers.
Yeah but we can already see that it doesn't work like that.
If you need to write a lot of code I guess, but that's really rare, like saying "I need to write a lot of laws. I need to write 50 new laws by Tuesday with at least 15000 words of new regulation to one-up my rival legislator who wrote 40 new laws last week"
If software engineers are more effective, I would expect there to be more software engineers. They’ll put out more and better code. More code means more engineers.
The contrary view is like saying gold miners are finding more gold, and it’s easier than ever, so we expect folks are going to leave town.
LLMs aren't the cause of mass layoffs and hiring freezes. The end of ZIRP, uncertainty in the macro and offshoring are the cause. AI is just something executives like to say recently when they do layoffs ("we'll be more efficient with cutting edge technology!").
I think there's real pressure from investors to show that some of your human costs will be going away.
After all most of those investors are deeply invested in AI technology already. At the valuation, they need to be able to show that it replaces human workers because that's the specific kind of greed that is driving the value of the stock.
And if you see your competition tighten their belt then you should tighten yours right? So without proof companies are acting like they can use a small number of human-ai hybrid workers. There's strong peer pressure to think that way as a direct result of AI
That's fair, but enterprises are often naive and prone to groupthink.
It was just a few years ago when automakers and rental car companies unanimously decided (has they had been told to decide) that COVID-19 would reduce demand for cars. They cut production, sold off fleets, and almost immediately found themselves unable to keep up with demand.
That AI is in a bubble akin to the crypto craze from a few years ago, and the valuation of these companies is divorced from their underlying business fundamentals
> If AI (terminology aside) replaces many professional software engineers and programmers like some of its fierce advocates say it would, wouldn't their potential customer base shrink?
This is such a good point. The best reply available to the AI hype-men would probably be that LLMs "democratize" coding and therefore that even more people will use IDEs in the future, but that sounds like BS to me -- not unlike AI/hype itself.
Steel man: Windsurf own the customer relationship. The models are just generic interchangeable services they use for processing.
Realistically: I don’t know how many users windsurf actually has and I never actually met anyone that uses them. Whereas Cursor AI took a huge percentage of the VS code users I know in real life.
Cursor purports $300M in annual recurring revenue (ARR) but stays silent on churn.
They made $25M from subscriptions one month, took that number, multiplied it by 12, arrived at $300M and everyone has been running with that line without ever asking what their churn looks like.
They could have churned $24M the next month, ask yourself why they are silent on churn if they are doing so well.
Venture capitalists aren't ignorant, their business revolves around knowing exactly what churn is. Cursor has raised $1 billion with a $9 billion valuation. VC's willing to put in that much money has looked at their data and knows what the retention rate is.
If their plan is to make their money back selling the company, then they don't care about revenue or retention rate. The company just need to look like it might be doing well.
No, venture capitalists aren't ignorant, but their goal also might not be to build and run a healthy company long term. It might be to turn a quick profit by selling a startup to another company.
"I'm skeptical about this VSCode fork commanding a $3 billion valuation".. Nothing to be skeptical about. The market has spoken. It was worth 3b to OpenAI. Companies arent worth a vague notion of what "value" someone in an armchair thinks they might be worth, they are worth what people are willing to pay, and OpenAI paid.
It's about popularity. OpenAI lost their monopoly now that there are many competitors so they're just trying to make a move to purchase "relevance". They're just trying to buy their way into the cool kids club, to remain relevant to at least a large number of kids.
You can't really make a living anymore being a furniture maker. Even semi-famous people in the industry have a hard time doing it. Only a few make enough to feed their family.
Making cabinets, etc.. sure. But woodworking has drastically changed, and maybe programming is changing that way, too.
Of course any tool that makes a job easier means that less people are needed to do the same job. If demand doesn’t change the supply will naturally shrink.
So hypothetically 1 man can cut wood but it takes him 2 days to do a big job. With a power saw it takes him half a day so his output on this section of the job is amplified by 4x. Any tool that makes his life more trivial increases his output and therefore increases the supply of the product without touching demand. With an over supply the system will naturally lower in supply by replacing carpenters.
This happens for anything and any tool that makes someone’s occupation easier. You have to think in aggregate. It may be the increase is imperceptible as it only increases the efficiency of a worker by 1 percent which is nothing but in aggregate that translates to a 1 percent reduction in the work force. Of course reality is more complicated than that but I hope the example shows you what I’m saying.
And it gets even more complicated than this too because increasing supply can also increase demand because the product becomes cheaper. Or demand may have already been astronomically high so the increase in supply only meets the demand.
In general if the product is in equilibrium of supply and demand and you increase the efficiency of the worker producing the supply then you will reduce worker population because the job doesn’t pay well enough anymore and people leave or less people join. The system slowly comes back to equilibrium or it can oscillate back and forth between over supply and undersupply as it’s basically a control system. This is what’s been happening with software for the past 3 decades.
The idea that the power saw didn’t replace a carpenter is flat out wrong. The story is much more complicated than that but the reality is that in general it did replace some carpenters just like how vibe coding for sure is replacing some software engineers.
You're right that it may have, and you're right that it's hard to quantify. But it is thoroughly possible that the increased productivity heightened demand. The ability to create more or better output for the same amount of labor can make some production feasible that wasn't before.
After all, in net, increased production has allowed us to have more. We aren't making do with the same amount of stuff and spending less. We're spending more, and receiving much more. That money is going to other people's pockets.
For comparison, JetBrains generates over $400 million in annual revenue and is valued around $7 billion. They've built proprietary technology and deep expertise in that market over decades.
If AI (terminology aside) replaces many professional software engineers and programmers like some of its fierce advocates say it would, wouldn't their potential customer base shrink?
Professionals typically drive enterprise revenue, while hobbyists—who might become the primary users—generally don't support the same business model or spending levels.
What am I missing here?