In this MonkCast Conversation, RedMonk analyst Kate Holterhoff talks with Evan You, the creator of Vue.js and Vite and founder of VoidZero, just days after VoidZero announced joining Cloudflare. Evan clarifies what the deal does and does not include, as well as the history of how it came about. He’s also candid about the motives that inspired VoidZero to look for partnership with a cloud provider. They also discuss the state of open source and frontend tooling in 2026. Can independent open source infrastructure survive today without getting bought by a bigger platform? Why Evan is still unconvinced that Vue needs a foundation. What is AI doing to the economics, and the craft, of building in the open?

Cloudflare is a RedMonk client, but this RedMonk conversation is unsponsored.

Links

• LinkedIn: Evan You
• X: @evanyou
• Evan You and Steve Faulkner, “VoidZero is joining Cloudflare,” Cloudflare blog, 3 June 2026.
• Evan You, “VoidZero is Joining Cloudflare,” VoidZero blog, 3 June 2026.

Transcript

Kate Holterhoff (00:04)
Hello, my name is Kate Holterhoff. I am an analyst at RedMonk and welcome to this MonkCast. My guest today is Evan You. He’s the founder and CEO at VoidZero and now on the open source team at Cloudflare. Evan, thanks so much for joining me.

Evan You (00:18)
My pleasure.

Kate Holterhoff (00:19)
I invited Evan to join me after chatting with him at Vueconf, which was conveniently in Atlanta this year and I was so jazzed to have him come on and just kind of talk Vue with me, talk frameworks, things like that. Little did I know that he was gonna have this, tremendous news coming out. So, I wanna begin by just saying, huge congrats to to the entire VoidZero team on the acquisition. You know, lots of interesting projects are coming out of Cloudflare. So, Evan, way to go. This is this is exciting.

Evan You (00:47)
Thank you. Yeah. yeah, everyone on the team is also excited. So yeah, I think we joined at a pretty good time.

Kate Holterhoff (00:48)
All right.

It’s it’s remarkable. Yeah, no, too cool. I’ve I’ve read all the blog posts that I could find on the subject, so I’m I’m hoping we can dig into some of the details here today. but for folks who are maybe less familiar with the front end community, with you know what you do, could you just introduce maybe, your background and and what it is that VoidZero does?

Evan You (01:16)
Sure. so I’ve been working in open source since like twenty thirteen, fourteen. so I was still at Google at that time and I started Vue.js as my side project. so Vue.js is a front-end JavaScript framework. so I worked on that.

And in 2016, I started doing open source full-time as an independent developer. So that’s like purely based on sponsorships and some you know partnerships with people who are making content on top of UJS and you know collaborating with people selling products to you know developers and they would sponsor the project or have different kind of you know small partnerships and which made it actually sustainable for me.

To just work on it full time for almost a decade, actually. Right. So in 2020, I created another project called Vite, which is a JavaScript build tool. and it got quite popular. so in 2023, I started a company called VoidZero, essentially trying to leverage the momentum and ecosystem around Vite and to expand that and build a full toolchain for JavaScript.

development and yeah the rest is what happened recently VoidZero was acquired by Cloudflare and now the whole team has joined yeah

Kate Holterhoff (02:46)
That is a clear and succinct history. I appreciate that. I I think it’s valuable because, many folks may have heard of the acquisition and and and be less familiar with the breadth of experience that Cloudflare is now bringing in. and so I guess just, worth noting, I invited you on here before you were an employee of Cloudflare. So, you’re you’re here on your own auspices, This is you you did this to yourself, Evan, is what I’m saying. You know, this is okay.

Evan You (03:12)
Yeah.

Kate Holterhoff (03:14)
all right, so exciting. I think f many folks in the front end community when they hear about this sort of an acquisition, they get a little nervous. And I feel like you addressed this very well in the blog post, but could you help us to understand like what is it that this is gonna mean for the Vue community, the Vite community? how are things gonna be changing from their perspective?

Evan You (03:38)
Mm-hmm.

So, first of all, for Vue, right? I also talked about this on X. so Vue is not part of the acquisition. Actually, Vue has never been associated with VoidZero in the first place. this is also a kind of a common thing I had to explain over and over again when I started the company. people are like, you know, what does Vue get out of VoiceZero? And I’m like, well, they are separate, right? Which is intentional. Vue is an independent project, it’s still my project, but

Kate Holterhoff (03:50)
Okay.

Evan You (04:09)
I intentionally have a very clear boundary between Vue and Voice0 of the company, right? So for example, the IP of you know, on the IP aspect, I actually signed the part of my IP of Vite into the company. but Vue has always stayed out of this whole thing. and similarly, in the in the acquisition, Vue was just never part of it. Actually, I have an explicit clause in my employment contract which says Vue is prior art, and it has nothing to do with my employment.

At Cloudflare at all. So yeah, so that I’ve always been really intentional in keeping the sort of independence of Vue because in a way it’s it’s a special project to me, right? And also my intention with VoidZero in the very f in the very beginning is to say, okay, this is going to be a company, this is going to be we’re gonna raise VC money, right? We’re potentially gonna sell it to other companies in the future, right? So I know this is potentially gonna happen down the road, so which is why I always kind of was really careful in not dragging Vue into it. in a way I like having the ability to sort of operate Vue in its own computer.

Kate Holterhoff (05:07)
Mm-hmm.

Evan You (05:20)
capacity without having to worry about, you know, having other stakeholders trying to get a piece out of it. You know.

Kate Holterhoff (05:26)
Yeah. Okay. What a unique situation too, because like we’ve had Daniel Roe on the podcast talking about when Nuxt was acquired by Vercel. I’ve talked to to folks on the the React foundation team, Seth Webster’s come on. So I I feel like there’s always this sort of nuance with these acquisitions. but yours is is so unique because you have so many important infrastructure and you know community type projects.

that you’re you’re balancing here. So I can imagine that the the lawyers were were having a hard time with that. and yeah and when we talked at Vueconf, you gave a great talk. It was it was like a you know an hour-long sort of discussion of of where does AI fit into this? Like what is the future of Vite, all these things. and a lot of benchmarks in there, and of course all the different projects. but one of my my takeaways too was that how much you had built

Evan You (05:59)
Ha ha ha.

Kate Holterhoff (06:21)
That was working with the Cloudflare ecosystem. So this feels like a very natural fit. so they, maybe I should have anticipated this. I am an analyst after all. so I guess you mentioned thinking about it between twenty twenty-four. How long it have you been, considering Cloudflare as like the the best home for this?

Evan You (06:25)
Mm-hmm.

specific for the acquisition it’s not too long ago. I guess the the initial idea started when I was at Cloudflare Connect in October 2025.

Kate Holterhoff (06:47)
Okay.

All right.

Evan You (06:58)
So I think that also was around the time we announced our Series A, but you know, these like funding announcements is always a bit delayed. We actually closed our series A earlier in the year. and I was at Cloudflare Connect because I think one of the things is Vite has always been this sort of entry point for web development, right? People start building apps with Vite and then

But for a very long time, people also think about Vite as just this like front end build tool. Like I they can use React or Vue or some other frameworks with it and they produce some static assets and that’s it. Right. when you think about full stat development, how do you get into the back end? Usually it involves another framework, right? So there are frameworks a lot of frameworks being built on top of Vite. There’s like Astral or Nuxt or

TanStack Start. but traditionally people kind of always, I guess before these Vite-based frameworks got more mature and popular, people always kind of look to Next.js, right? And when you look at Next.js, naturally there’s this Vercel sort of story where Next.js helped Vercel get off the ground and you know get all the users. so

From my perspective, like when we’re working on at VoidZero, we’ve obviously tried really hard to see if there is a way to just monetize on top of the tooling itself. But as you can imagine, that’s actually really hard. so down the road, we’ve always been thought, okay, like, you know, eventually we may have to go into just thinking about how we can leverage Vite and build something that’s tied into the server compart compartments. Right? Because like monetizing on tooling or front end alone is extremely difficult.

But the moment you have a server running somewhere, right? You are running compute for people. That makes it easier for people to pay you. So I guess down the road we always felt like if you know just like pure tooling-based monetization doesn’t really work out, we always have that around, that option. And in fact, Void, which is the project we kind of you know launched right before the acquisition, is our exact attempt at that. So we’ve been thinking about potentially going.

Going into a cloud offering, I guess, like for a very long time. One of the reasons we didn’t do it is because we just had so much work to do before that. And also we were kind of trying to see if, you know, just the ecosystem, how does it mature? Obviously, I’ve been doing a lot of research on my own and just thinking about if I were to build a full-stack solution.

That’s kind of serves as an extension of Vite, right? What it would look like and where should it run. in a way, that it ties into the question of like where should it run? And usually people would say, okay, it needs to run everywhere. You need to be vendor agnostic. which makes sense for a tool like Vite, right? But at the same time, I think

The thing with vendor agnostic solutions is that you are always constrained by the lowest common denominator. because the same thing needs to work everywhere. Which means, like if a platform offers something pretty unique, but you cannot replicate it on other vendors, then you can’t can’t support it, right? which is a pity. So and unfortunately, no one has really explored that, right? So when I look at Cloudflare, I’ve always, you know.

That for conflict specifically, I’ve always been like a fan of the primitives they’re offering. you know, it’s very effortlessly scaling globally, and they have they added over the years they’ve added D1, R2, you know, KV. a lot of the primitives you typically need for an application, but

If you’ve used Cloudflare, you probably know that the the experience of stitching these things together and get your code to actually run on Cloudflare hasn’t always been a super smooth experience, right? So I just saw the sort of, you know, it’s it’s a space I’m interested in. It’s also I see some sort of opportunity where we can potentially leverage feeds power to make that experience smoother. And we’ve already had some collaboration with Cloudflare prior to the acquisition as well.

because because like i said Cloudflare is a pretty unique runtime it has a lot of its own quirks and when people try to do local development they don’t have the actual Cloudflare system running locally right so it’s very different from say if you’re deploying to a pure node.js hosting platform

You’re running Node.js locally, you’re also running No Node.js on the server. So you expect your code to behave exactly the same. But that is not the case for Cloudflare for for a very long time, right? like maybe a few years ago, if you were to deploy it to Cloudflare.

you’d actually be running your code locally in Node.js, and then the deployed code runs in Cloudflare workers, and there are different runtimes. There are a lot of behavior differences. That’s just not consistent. And you only find out after you deploy. That is not ideal at all. So what we kind of work with Cloudflare on was this thing called environment API in Vite. It’s in a set of API that allows you to basically take the user’s code, even

Even if the user is starting their local development process in Node.js, which is Vite runs in by default.

Vite can actually take their code and then put it inside a separate, a different runtime, which is, for example, Cloudflare’s runtime. It’s called workerd. So we can put you take the user’s code, transform it through Vite’s pipeline, bundle it on the fly, and make it run in workerd locally on the user’s machines. Now the behavior will be consistent between the local development process and you know the remote, the deployed version that runs actually on Cloudflare.

kind of solves a big you know gap in the development experience for Cloudflare and that’s a also part of where the synergy came from as Cloudflare saw like okay Vite can actually help us s solve this problem and for me I was just like looking at like hey we can solve a lot of problems Cloudflare have we can also give Vite a natural sort of step up. Like for a Vite user they can now say I want to build a full stack application. I can just do things like this and it just

run on Cloudflare without having to go through all the extra trouble. so I started I guess like we we we didn’t like explicitly talk about this until I I came out with void right so there has been a bit of interest on both sides when when we talked at Cloudflare connecting October right but it wasn’t

It wasn’t like, hey, let’s just like start this whole like acquisition thing. It was more like there could be more opportunities for collaboration here. And I think when Void came out, it just became super obvious that, you know, the synergy is just like on our side we have, you know, built this thing on on Cloudflare that shows them, hey, like there are great things that can be done here and on Cloudflare side they’ve always been thought thought thinking of Vite as this

Kate Holterhoff (14:09)
Yeah.

Evan You (14:34)
Think that A helped them with local DX a lot and B can potentially bring more users onto Cloudflare.

Kate Holterhoff (14:42)
Yeah.

I appreciate that you have brought in that the subject of like how Cloudflare is gonna be leveraging Vite. Because when I read that press release, I assumed it was gonna be more like, no, we’re just, you know, we’re adding this to our what our portfolio here, and that this is gonna be more like, benefiting, you that you’ll you’ll take advantage of all Cloudflare’s resources.

But no, it comes across very clearly what you’re saying, that they’re going to be building on top of Vite in these significant ways. and it actually kind of reminded me a little bit of how Anthropic has w when they acquired Bun, just in the sense of like, these are open source projects, but there’s something to bringing that team in-house, having the skill set on hand and also just that deep knowledge of like

how to leverage these tools and to to take advantage of them to to their fullest. so it I I mean it speaks a lot to maybe Cloudflare’s own strategies and and what they’re looking at in the future. I mean needless to say AI takes up a big part of their announcement as well. so yeah, it there was a lot of takeaways there. I don’t know, did did you want to respond to any of that before I move on?

Evan You (15:56)
Yeah, I guess Times back it I want to talk a bit about Bun like you know like companies acquiring open source, right? it’s like there’s always this concern about like will this thing stay neutral, will this sta thing stay free? I think I do want to emphasize Cloudflare has made very, very serious commitments here to ensure that, you know, like we obviously the purpose of the acquisition is to, you know, leverage V and try

Kate Holterhoff (16:00)
Yeah.

Mm-hmm.

Evan You (16:26)
to get more people to use Cloudflare, right? That makes total sense. But we want to do it in a way that does not distort the incentives

Of the project itself, right? The goal of the project is still to be the best tool possible, right? And in order to get people onto Cloudflare, we do that by keep making Vite great and just making the specific part of Vite and Cloudflare integration smoother than before. But it should not we should not do it at the expense of making it worse for the other platforms. So that is super important for us. Because in order to keep Vite great and keep people trusting Vite, we need to keep it, you know, usable and beneficial.

Kate Holterhoff (16:56)
Mm.

Evan You (17:05)
everyone else even if they’re not using Cloudflare.

Kate Holterhoff (17:08)
Yeah. Okay. That makes sense. And so I but before we move on from the monetization issue, ’cause you know, obviously that’s that’s really important here. I’m interested, did you consider any other business models before you moved into, thinking through acquisition? For instance, having a a sort of mixed licensing mode or, the void deployment platform. how serious were you with those initiatives before deciding that acquisition was the way to go.

Evan You (17:36)
it was obviously very serious. we are a VC funded company after all, and we we know, you know, we were under pressure to find a actual business model. so Vite+ was the original thing that we were trying to see if the the mixed licensing model

Kate Holterhoff (17:39)
Okay.

Right.

Evan You (17:55)
So for context, we we built this thing called Vite+, which is a essentially a combined unify tool that combines Vite, Vitest, Rolldown, OXC, all the things that we’ve worked on and more, right? It contains something that’s unique to Vite+ itself. For example, we have the task runner system with automatic inferred caching based on file access. we have

Kate Holterhoff (18:09)
Mm-hmm.

Evan You (18:22)
It bakes in Node.js version management and patch manager management. So it’s it’s almost a one-stop shop for your whole JavaScript TypeScript development workflows. and we think that actually we we’re still pretty, you know, confident that it’s probably the best thing for your agent to use when it comes to to TypeScript development. and we thought it would be valuable enough so that we could

Is it so we well the license model we tried is kind of interesting because we made it free for for most individual users, right? And the license model we tried is companies above a certain revenue threshold would have to pay for it.

Kate Holterhoff (19:02)
Yeah.

Evan You (19:04)
Right, so so we were testing the waters with what that threshold should be, and we got a very long list of companies interested in it. We actually got to, you know, conversations of people actually are like, Hey, we want to actually just pay for this so we can use it. Like, what is you know, what what exactly how much do we pay you? and we got into those conversations, but when we got into the actual details of like what they actually want, they started asking about support and they started asking about like you know.

how many hours of support do we get each month? And if we run into issues, who do we talk to? And we realized, I guess, the expectation with selling software like this is because it’s not it’s not like a hosted service where as long as things keep running, people are happy. using tooling is you know, inevitably they will have to they will run into all kinds of issues. it’s kind of like

you get so we already get a lot of GitHub issues. We already get a lot of bug reports. And now once they become paying customers, we kind of have to prioritize this and we have to have people just on Slack sitting all day answering questions pretty much. And we have to build a whole support team.

And we realize like the scalability of this isn’t ideal, because we don’t want to turn into a support, you know, company. Like I mean, no disrespect to Red Hat, but like we don’t want to have to like have a team of, you know, like 50 like so our core team is just like less than fewer than 20 people, but we don’t want to have to build another like 50 people solution engineer team just to answer questions all day. that is just not the business I sign up for. so

So we then we started thinking about void, like I said, like you know, selling hosting is has always been the kind of thing like we know we can do if we really want to, right? the reason we didn’t get it into too early is like we’re like, let’s explore what we can do with Vite+ and quickly we realize okay, like it’s not the it’s not the thing we really want to do, right? Because if you’re in a company

It’s always kind of tough balancing like, you know, how do you actually make money and what do you really try what are you trying really trying to achieve? and I think in a way, having

Having good investors is nice in that way because our investors really didn’t like pressure us so much in like they were not short-sighted enough to force us to just like you you need to actually start having revenue like by year two or something, right? So so we got the opportunity to try things and then we got the opportunity to explore what we can do with void and

Kate Holterhoff (21:51)
Mm.

Evan You (22:01)
And for Void specifically, I guess it definitely helped the conversation, the acquisition conversation, because it just made things so obvious. but when we realized, so for Void specifically, it’s it’s a whole the reason we didn’t get into hosting too early is because hosting is a whole domain of its own, right? Like once you get into the hosting business, there’s this whole new category of things like how do you design billing how do you, you know.

Some companies may want SLAs, you know you get to get through like all the security, data retention, a lot of problems that we did not have to deal with before, and now we have to deal with it, right? So as we are the as we are sort of still in the private beta phase of Void, we started to actually working on these problems and we’re like, well, this probably is gonna take quite a bit longer than we hoped before we can actually say this is production ready and start making real revenue from it.

But then again, in some way, it’s great that we can build this thing and eventually make money. But also, like, you know, we feel like a lot of our passion is in the tooling itself. and now we have this whole hosting platform thing. And we need to hire more people to work on hosting. I have to split my attention between tooling and hosting.

It’s you know, deep down we’re like we wish we could just keep focused on tooling, right? And so the acquisition actually just gave us that opportunity because look, we’re building on top of Cloudflare. Cloudflare is already offering hosting. It’s like compute, you know, all the

Kate Holterhoff (23:30)
Mm-hmm.

Evan You (23:41)
It’s a cloud platform, right? And the cloud platform wants to support us and we can keep working on tooling. And our tooling will benefit the the cloud platform. It’s it’s basically, you know, allows us to kind of skip that phase and directly jump into what we actually want to do.

Kate Holterhoff (23:57)
okay, well let’s stay on that subject of like deployment vendors, like these cloud platforms. so Cloudflare obviously fits in that category, but some other ones include, you know, Vercel, Netlify, there’s many others. and it is no secret that Cloudflare and Vercel have had a little beef. I as I’m thinking back on Connect, I remember Sunil Pai showed me a blog post where there was some

Some static going on there. so no need to to be quoting tweets here, but but we’re we’re we’re all familiar. I know I know you’re you’re chronically online, Evan. so talk to me about how the community should feel about that, especially folks who maybe use Vercel Like should folks who use projects like Nuxt and Svelte should they be concerned at all about support for Vite or any of the VoidZero products.

Evan You (24:50)
Mm-hmm. Yeah, I think honestly, I think the beef is it’s just like it’s a social media thing. like I c I’m pretty sure like they are

Kate Holterhoff (24:59)
Okay, fair enough.

Evan You (25:05)
to some people it’s on a personal level, but I think at the business level, obviously there are they are competing businesses, right? They are trying to compete for the same sort of positioning as like this is the cloud for agents, the cloud for the future of web deployments. which I think there is so I think there are basically three parts. A, there’s this like personal beef stuff, which I don’t think matters in the grand scheme of things.

And then there’s the actual business competition, which I think is it is what it is, right? companies gotta do business. And then there’s the people who are actually working on these things, right? which is very different from the beef people are perceiving. actually, you know, the the people who joined Vercel as part of the Nuxt acquisition. In fact, one of the Vue Core team members, Eduardo, who’s probably one of the oldest members of the Vue team, I’ve been personally sponsoring him for over

Kate Holterhoff (26:01)
Wow.

Evan You (26:03)
all these years and he just recently joined for cell and is actually really happy about it and he’s like you can stop sponsoring me now because I get paid pretty well by Vercel and I can still you know spend some spare time working on Vue which is great so I’m still good friends with the people on the Nuxt team with Eduardo there’s really nothing that changes the dynamics of which is also a perk of me keeping Vue independent is we can still keep collaborating on these things on the side.

it doesn’t really get affected by how you know some high-level people in these companies kinda hate each other. it doesn’t really affect us, right?

Kate Holterhoff (26:40)
Yeah.

Well put.

Okay.

Evan You (26:44)
Yeah, and and I think it’s it’s also like even even at the open source parts, I think generally people working even on the in these two companies, even they’re working on competing products, they have a lot of respect for each other, right? ‘Cause you know, it’s it’s it’s it’s just a matter of fact that Vercel is shipping a lot of great open source and Cloudflare is shipping a lot of great infra products and there are things we can learn from each other, right? and we have people who previously worked at Vercel now working at Cloudflare.

vice versa. So there’s a lot of cross-pollination. And I think for the users, I generally think it’s actually better that there are companies, cap competent companies competing in this space. It’s kind of like you know similar to how I think the the AI space would be worse off if Anthropic and OpenAI were not competing, right? Because it’s the competition that forces them to kind of just keeping up the pace and trying to one up each other. I think it’s

And similarly, you know, in the web development space, in the in the agent hosting space, you know, you gotta have competition to have people just keep innovating and coming out with great ideas and better implementations and users end up benefiting because you know because like

having choices also forces these companies to avoid doing you know blatant vendor locking because if you do that people see that and people are not stupid they will be like if you want to lock me in I’m just gonna not gonna be using you for the first in the first place right so I think overall you know this kind of competition just benefits end users.

Kate Holterhoff (28:09)
Mm-hmm.

Yeah.

Yeah, a and as part of that, I know that the press release talks about this, one million dollar independent fund as part of the pledge to neutrality. Do you have any insight into who’s gonna be seeing any of that money? ‘Cause divided by all the maintainers isn’t gonna be much. but, it’s a nice gesture.

Evan You (28:39)
Yeah, we we actually already have part of the part of the deal considerations flowing to some of the long term contributors because they we actually granted some of the contributors some options in VoidZero.

Kate Holterhoff (28:48)
nice.

Evan You (28:52)
So they they already, you know, there are already some money flowing there, and this one million actually is like extra on top of it. And this is gonna be openly administrated by the Vite core team. so Vite core team consists of me and Safi, who is a VoidZero employee, but there are also four other people who are not part of VoidZero or Cloudflare. Right? So it’s going to be just this like very

Kate Holterhoff (29:06)
Okay.

Evan You (29:22)
kind of open ledger thing if we were to disperse the money it’s going to be just like public records. Who gets what? You know.

Kate Holterhoff (29:28)
Okay. Is it an ongoing fund? Is there gonna be like a million dollars every year? It’s just like they’re this is the payout as like a Okay.

Evan You (29:34)
Right now it’s a one time. we probably it’ll probably take some work like for us to like you know even make use of this properly. so yeah. So Matias, who is probably the one of the

Kate Holterhoff (29:44)
All right.

Evan You (29:52)
He is actually the first person I got onto the Vite team. and he’s been with us through all this time. So he’s the com our community per community person. He is just like the helped us get things off the ground, talk to all the people, making sure they are aligned. he knows everyone and he is also really well respected. So gonna, you know

Kate Holterhoff (29:56)
Wow.

Evan You (30:14)
I trust him in just making good use of this. obviously I’m going to be sort of involved in the process to say this looks good, but I kinda want to sort of I don’t want to be the person to say you get this much, you get that much, right? This has to be sort of a team based decision.

Kate Holterhoff (30:23)
Mm-hmm.

Okay, okay. So let’s take a step back maybe and talk about the lessons from this experience because Astro joined Cloudflare you know, what is it earlier this year? Now Vite. so what sort of advice would you give to folks in that are maintainers or that are have built these projects, that are open source and are maybe struggling with the monetization issue that you’ve struggled with? you know, what sort of advice would you give to them? And and frankly,

Just like from the state of software today, what would you say we’ve learned? Like what is it viable to have standalone open source infrastructure in 2026 without being absorbed into these like infra giants? you know, is that really the only end game possible? Like and and maybe have things even shifted? Is that how it’s always been? Like this was always

You know, th this has always been a problem. I don’t know. So yeah, lessons that you would share with folks who are in similar situations and like, where are we today?

Evan You (31:37)
Mm-hmm. That’s a sharp question. and so I think for this question there are two different ways to think about it, right? So I kind of have experienced two both sides of it, right? I still keep Vue independent, and Vue is sustainable, even without the acquisition money that I

Kate Holterhoff (31:48)
Okay.

Mm-hmm.

Evan You (32:01)
I personally so now I after the acquisition I’m planning to personally kind of you know put a bit more to support other people working on Vue. But even without that, Vue has always been sustainable. Like we have Open Collective, we have GitHub sponsors, we have historically we had a bigger Patreon sort of income, but like it’s much smaller now, but there’s still something going on.

that we allow us to support maybe you know two full-time people actually just working on Vue and nothing else. but Vue has historically even like so Vue given Vue’s usage scale right it’s downloaded like I don’t know like

close to 10 million times a week now. so it’s like the second most popular JavaScript framework in the world. But the amount of sponsorship we get, it’s sustainable, but it’s definitely not not, you know, not substantial by any means compared to the scale of impact it has created. Right. So

But even that Vue has been an outlier for independent open source. I’ve seen many projects like Babel, which has been critical in the JavaScript ecosystem. It has historically always struggled with funding. and then we have so if you are running a an an open source project purely out of out of your passion.

Kate Holterhoff (33:12)
Mm.

Evan You (33:36)
nowadays I think it is actually really difficult, right? before I started VoidZero, people have always come to me to say, Hey, how can I make my open source project as sustainable as Vue? I always feel like

You know, A the thing I do with Vue is very hard to replicate because every open source project is different. It’s kinda like every startup is different. You gotta find your own way of monetization, your own way of you know, your own business model. And open source is similar.

Vue is unique in the sense that it has a big reach, it is a JavaScript framework, people interact with the documentation a lot, it has great exposure to end developers, right? But if you’re a a more tooling infrastructure layer thing, like for example, Vite, right? Vite actually quickly has more download numbers than Vue does, but like Vue’s, Vite sponsorship numbers has always been lower than Vue’s because it just you know when people use Vite, they treat it

As a tool. They install it, they get it configured, they get it working, and they forget about it. Right? They they love it when when it works, but once it starts working, they don’t think about it as much. Unlike a framework, when they write code every day, they would look at the docs and they have deeper engagement and then the sponsorships become more meaningful. for

So so companies are more likely to sponsor frameworks compared to tools. So there are a lot of these intricacies between different kinds of props open source projects. So it’s very hard to give people very generic advice on how to monetize your open source. So on the other hand, we see a lot of

Kate Holterhoff (35:15)
Yeah.

Evan You (35:21)
you know, open source with different like motives, right? There are open source projects where you start just because you want to do it. Like I started Vue and Vite just because I want to do these things. I didn’t really think about commercialization or monetization when I started these projects. so it’s more like passion driven. And then I try to figure things out along the way, right?

The way Vue sponsorship model works, like it just organically evolved. I didn’t really try to force it into that way or try to like design it from that way in the beginning. but then we also see there are a lot of companies or startups nowadays where they’re doing open source, right? But open source is only part of the whole commercialization strategy. It’s a means to an end. the whole idea is to build a successful business, and open source is just a strategy.

that they happen to pick to to help them achieve that goal. Right. So I think the incentives is very different. so I would say the the first category, the passion driven ones, it’s getting harder and harder, right? especially with AI. Because the thing is the the cost of A, like you no longer get the same level of

the the the joy of craft when you work on these new ideas. so when I worked on Vue, I I would keep, you know, polishing these little details over weeks, months, and I I was happy just working on it because, you know, it gave me this sense of like sort of I’m I’m doing something that’s really great, I’m proud of. But I think AI nowadays sort of A it kind of takes away some of that joy because it’s so easy to just come up with an idea and like

Tap a prompt and you get something working in like 20 minutes. it and the ideas are also much easier to copy because like you can open source something and the Next day someone would fork it with AI and add a bunch of features and say, I made a better clone of yours stuff. so yeah, it it’s it’s tough. I think

Kate Holterhoff (37:13)
Right. Yeah.

Evan You (37:34)
I don’t really know what to make of it at this at this point, because like if if you’re just starting out a new passion driven open source project, I think like it’s definitely tougher than what we got back then. and monetization would even become more of a challenge. And and even with commercial open source, there’s been this kind of like, you know,

There are companies who are just closing off issues because people are sending spam PRs with AI. And there are people just go went straight to closed source because they worried about people just you know forking their thing with AI and just like be a become a competitor overnight. so the dynamics is changing. and there are definitely a lot of new problems we need to think of.

when it when it comes to the interaction between AI and open source. And honestly, I don’t have clear ideas or answers to these yet. I’m concerned, but at the same time, I think what I do want to encourage people to do is to still keep you know keep the the open source sort of etiquette or

Just like be be good actors in the space, right? Still sort of respect what original ideas that people come up with. Don’t just like take people’s stuff and fork it just because you can with AI. Right? try to collaborate. Try to still like be pull your efforts together to make great things together rather than like everyone can just do their own version because AI is gonna write all the code. I don’t I don’t I That’s not the future I want to see.

Kate Holterhoff (39:25)
Yeah. Agreed. I mean on that note, I gotta bring up, you know, Cloudflare is actually I wouldn’t say notorious, but they they have a history now of of what, creating vibe coded versions of front end things. Like so they they just did most recently a a a a WordPress clone as well as a Next clone. So I’m I w was there a part of you that was worried they were gonna do a Vite clone?

Evan You (39:50)
Not really. I think the Next one is is controversial, but there’s been a long history of people wanting to deploy NextJS on Cloudflare and running into all sorts of issues. and in a way I think the vinext thing is really just came out of the frustration in trying to support all the edge cases have like trying people trying to run Next.js on Cloudflare. And they were because the the fork on vinext

Kate Holterhoff (40:04)
Mm-hmm.

Evan You (40:20)
is actually Next running on Vite. and it runs much better on Cloudflare because it’s based on Vite. so that’s the that’s the actual reason. It’s not like forking just because they hate each other, to be honest. so there are there are technical reasons why they why they try to do it. And vinext is in fact also used in a pretty prominent project

Kate Holterhoff (40:29)
There you go.

Evan You (40:46)
Actually open OpenAI sites, as far as I know, is deployed on Cloudflare and running on vinext.

Kate Holterhoff (40:55)
Well, I didn’t know that. That is interesting. It seemed like a hobby project, kind of like a look what like a cool look what we did. I didn’t realize anyone had actually deployed it. Because y is that I think it was TJ VanToll who is on his own podcast, the Front End Fire one, was actually talking about trying to use it as a drop in replacement and because there’s no like documentation or there wasn’t, I d I don’t know what what it looks like today, he wasn’t able to get it working.

Evan You (40:58)
Yeah.

Kate Holterhoff (41:21)
Now, I don’t know how much time he spent on it, maybe, you know, I’m sure there’s extenuating circumstances here, but I this is the first time I’ve heard of it actually being used. So this is this is a good proof point. I this is interesting.

Evan You (41:30)
Yeah. The use case, yeah.

There are there are a lot of different use cases for these things, right? Because I honestly wouldn’t recommend people trying to say take an existing production level Next.js app and try to migrate it over.

There’s really little point in doing that. I think the idea is you can quickly spin up new projects that’s still using Next.js style code, right? But make sure that it can actually run smoothly on Cloudflare without problems. That I think that is the main incentive. Steve, if you see this, correct me if I’m wrong. Steve is the the one who worked on these things. and

Kate Holterhoff (41:45)
Yeah.

I see.

Evan You (42:13)
Yeah, and I think the in a lot of ways EmDash isn’t even a fork in a sense. I personally do think the whole there are a lot of like public communication around these around vinext and EmDash can be improved. Like I’ve actually talked to Dane about this. I think I think like we acknowledge there are definitely like you know certain ways we can communicate these things better.

Kate Holterhoff (42:30)
Yeah.

Evan You (42:41)
like for example for EmDash it’s not even a fork per se. It’s like it’s written in a completely different language. it does the same functionality, but it’s not like we take it’s not like they took the PHP codebase and tell the AI to say, hey, like rewrite this in TypeScript. it’s a pretty different codebase. it just does the same thing that WordPress does. and in a way

I don’t know if like calling it the spiritual successor of WordPress really helped, because like, you know, obviously a lot of people took issue with that. so yeah. Honestly, I’m not too involved with that project, so I probably should reserve my comments here. so I only know enough about the technical implementation of it to know that it’s not just a slop fork. Yeah.

Kate Holterhoff (43:32)
Yeah. Okay. Well that’s fair enough. I I won’t force you to speculate on on what’s going on there. But it obviously got a lot of press. And I think it just also signals something about what Cloudflare you know, th their strategic approach to front-end top-of-the-stack technologies. I mean, it’s just it’s been interesting to to follow that because I I I think it kind of is is a little surprising. when we think about what Cloudflare

was known for in the past, to be, the CDNs, things like that, more infrastructure level. I mean I actually, as a former front-end engineer, I love it. I’m like, well look, look how important we are. This is it. Everybody knows. We’re all paying attention to the front end now. This is the this is where the cool stuff is happening. but of course, Evan, you’ve already you’ve always known that, so you know, I don’t have to tell you. okay, so I want to hear a little bit about

Evan You (44:04)
Yeah.

Kate Holterhoff (44:24)
your thoughts on the open source situation because that is th was one of the things that we discussed actually at Vueconf. so talk to me about your overarching thoughts on the state of foundations today because is famously independent. what I I’ve heard it characterized as a benevolent dictatorship, as many great open source projects are, right? have you considered

Putting Vue into a foundation, I mean now seems like kind of a pivotal point. You said it was w written into your contract that it’s independent. many folks have reservations about foundations, understandably. I guess I’m just interested in, where do you sit on the foundation question?

Evan You (44:56)
Mm-hmm.

Yeah, hist yes, we’ve been asked whether we want to put Vue into a foundation. I guess I just really didn’t see the benefit back then, or still don’t really quite see the the full benefit of it. part of it is because There are two

types of situations where where I see projects going into foundations. A is the project is so big, so critical, right? People put it in a foundation with many big companies back in it and they’re like, this is too important to to be left alone. So we gotta just like support it all together. Which is like Linux or React.

So, but like I guess like for Vue for a very long time it’s just like in this sort of like limbo where it’s kind of big enough where it’s like people kind of think it deserves a good sort of sustainably model or like people we need to take care of it, but it’s also not big enough for multiple big companies to say, hey, we need to put money into this foundation to keep Vue going. There just isn’t such incentives. So

And if we want to get enough companies to be interested in this, that’s gonna be a lot of work on itself. And but and for for me personally, it has been sustainable already, right? So we just didn’t feel the urge to go out of our way to like build up the foundation, then try to like pitch all these companies to pull money into it. so it it just didn’t really make sense for us.

Kate Holterhoff (46:47)
Mm-hmm.

Evan You (46:48)
And my personal take on foundations, I guess like I don’t want to be I guess historically I just don’t really get how like what actually foundation would help for a project like Vue specifically. I can only speak from my own experience, right?

For some some other projects maybe they are not they’re not as sustainable as Vue on themselves. It definitely helps for them to just look for a home where they can be like, hey, take care of this. but for Vue, it’s like when we when we were asked do you want to join a foundation, I’m like, what do we get out of it? And they’re like, IP protection will help with this and that. But a lot of that for us is just sort of I guess

not critical. and yeah I just through those conversations I just never really found convincing reasons for us to say we should put Vue in a foundation. and I think even after this I also thought about you know setting up a foundation myself so that it could be

I I actually thought about this, right? I looked into how to do it, but like I obviously getting certified as five one C three is is a very complicated process and I just never had the time or energy to follow it through. so what I’m doing now is kind of like we have we’re probably gonna have an entity that is still dedicated to supporting Vue, but whether it’s going to be in the form of foundation or just a, you know, normal business,

Like to be honest, I just don’t really don’t really think it’s gonna make a fundamental difference for Vue specifically at this point, as long as w we can make it sustainable. maybe in the a few years down the road when I really want to say I’m I’m going to be retiring. I don’t you know, I need someone else to take care of it. Maybe that’s when the time I’ll think about putting it into a foundation. But for now, I just feel like, you know, I don’t

I don’t think it adds real value to to Vue specifically. I can’t really speak for other projects.

Kate Holterhoff (49:07)
Okay. and then I just I I gotta ask you about some the benchmarks and Rolldown. And I have done a little bit of research into some of the bundlers and I’m curious about how you’re thinking through the single page applications and how they relate to just I guess this arms race right now in terms of like Rolldown Rspack,

Evan You (49:31)
Mm-hmm.

Kate Holterhoff (49:34)
TurboPack, all of these different And and what is your so okay, so I did a number of interviews as part of my research on this. I didn’t get a chance to interview you, so this is my opportunity. I gotta I gotta do this. So talked with Tobias Koppers, creator of Webpack, and now TurboPack. I talked with Zack Jackson, who’s an engineer at ByteDance, he works on the Rspack team. And s I you know, I got some spicy takes during these interviews, but

Evan You (49:37)
Mm-hmm.

Mm-hmm.

Mm-hmm. Yep.

Mm-hmm.

Kate Holterhoff (50:01)
Both of them pointed me towards artifact size as being a big component of the reason that we have this arms race. Because at the end of the day, it’s hard for engineers to even keep up, right? It’s 10 times faster, it’s 50 times faster, you know. I it’s great. But I I do think that they’re changing all the time. Developers get this sort of benchmark fatigue. AI has only made it worse. And so I’m I’m interested to hear your sense of

Evan You (50:08)
Mm-hmm.

Kate Holterhoff (50:28)
this because I I I mean I I I’ll say like Zack Jackson had the maybe the spiciest that like he it’s but the Bundler war is really about artifact size and that’s the dirty secret. I love I love that term, dirty secret. So do you agree with that that it’s really just about the size or have we moved past this? Is the spa wars, you know, is that overblown? It’s not really about JavaScript size being pulled down onto devices like are we wh where do you sit on all that?

Evan You (50:29)
Mm-hmm.

Yeah, I so I agree. Like artifact quality is actually important. And every bundler is more or less trying to balance between raw performance versus cause like there are two different metrics, right? One is how fast you can build your app, the other is how fast your end user loads your app, right? So for engineers working on a project,

Bundling performance helps the developers because you can debug, you can build and debug your app faster, you can run your CI faster, you can run your tests faster, right? There are real benefits there, but for your end users, outer fact size is what determines how fast your app can load, how well you can split your bundles so they get more efficient page loads. You know, they’re able to load just a fraction of the code to be able to get the page load faster. So both are important, but they’re not.

mutually exclusive right so I think historically webpack has been this really really powerful extremely powerful thing and it has It gives you all these, you know, knobs and levers so that you can fine-tweak your artifact output so that so that you get the best possible like end user performance at the expense of being the architecture being a bit more complicated, and therefore the performance could also suffer because you know all the additional complexity it has it has to handle. so when we started with so

It could also I or would it also involve some other, you know, notable solutions like esbuild, which is made by Evan Wallace, who’s the CTO of Figma. esbuild is probably the first bundler that really kind of started this arms race when it, you know, its benchmark basically showed like esbuild is a hundred times faster than Webpack plus Terser, which is the standard combination back then, right?

Kate Holterhoff (52:50)
Wow.

Yeah.

Evan You (52:52)
So that was when like, you know, when that came out, people realized, okay, like we can actually have fast bundlers, right? But it is true that the artifact quality of esbuild is not as great as, say, Webpack Plus Tursor, right? historically, esbuild has larger bundles, the minification quality is worse, and it doesn’t really have any sort of code splitting control. It only has one automatic code-splitting mode, and that’s it. So you can’t really like do fine-tweaking to your

needs so if you use esbuild to bundle a large user-facing application your user will actually have worse performance like they will load your app slower right so that’s the trade-off right so when we started with with Vite we initially start used Rollup right and then we started working on Rolldown which is our own Rust implementation that is now powering Vite 8 and above and for Rolldown it really kind of takes time for the thing to mature because

you gotta first get things working, then you get into these deep optimizations, right? So the nice thing about Rolldown is like when we got the first basic functionality working we know okay like it’s like three times faster than Rspack and we’re both written in Rust right

So that’s kind of like where, you know, we’re we’re still more than twice as fast than than Rspack nowadays. and over time we’ve been adding. So so the interesting thing is in the beginning, Rodon is indeed producing noticeably bigger bundles. And it doesn’t have as a powerful sort of control over code splitting as Rspack does. It is true, right? So that’s why we have been consistently trying to improve on that front. but the nice thing is we’ve been able to add

Add all of these features. We’ve had so we’ve implemented better minification, we’ve implemented better tree shaking, we’ve implemented better code splitting over time without degrading the performance. So that is the important thing, right? So so overall we’ve been catching up in terms of artifact quality. And in fact, I think we’re really, really close to Rspack nowadays.

So the capability is pretty much equivalent. At the same time, Rspack has been making great improvements in terms of performance, right? So they’ve always been this like more feature-rich because they they are effectively a one-to-one port of Webpack, which has all the things that allows you to fine-tweak all the things, right? So initially they started a bit slower, but they’ve been tweaking the performance. So they’re getting a bit faster over time. So we are basically starting with this like much faster, but not as feature-rich.

kind of starting state and we are trying to stay as fast but we’re adding more features.

And they’re coming up from this like really feature rich but slower kind of place, but they’re optimizing the performance to try to catch up. So I think, you know, at this stage they are still still slightly better in terms of like overall artifact quality. I’ll give them that. but it’s really, really close now. And at the same time, they’ve been getting faster over time, but we’re still you know, comfortably faster. We’re like still more than twice as fast than our S Pack. So I think this is again a case of you know healthy competition where

Kate Holterhoff (55:39)
Hmm.

Evan You (56:09)
In fact, like four of our current team members came from they worked on Rspack at some point. So, you know, we we have like people who were previously worked at Bandance, now currently working were at VoidZero, now actually working at Cloudflare on Rolldown. So it’s a lot of like people that we know each other and we kind of you know have a lot of mutual respect as well.

Kate Holterhoff (56:36)
Yeah. Okay. That’s exciting. a and now you’re you’re able to leverage all of the engineering talent at Cloudflare to to keep working on this problem. I mean, yeah, Zack really pointed to the the fact that he he also mentioned Terser and that a lot of that code doesn’t run. So it sounds like there is a lot of opportunity to to try to optimize and make sure that these things are, I don’t know, as stripped down as possible. So and you know, using

I mean everything’s on Rust like the big I think that was like step one. Everything’s on Rust, or Go. and and after that it’s like just these these tweaks that you’re mentioning. do you find that developers are overwhelmed by all the these benchmarks and like I I feel that way, but maybe I’m projecting. Like, do you do you think that they

Evan You (57:02)
Yeah, yeah.

I think we are I think what we the the reason we’re d we did all this work at VoiceZero is to eventually try to you know, converge give people a sort of like this is, you know, the default starting point, so you stop worrying about all the stuff. I think it’s still

Good to have choices. I think Rspack is particularly useful for people who are coming from a legacy webpack webpack application, and they’re like, we just use all these advanced webpack features and we can’t just like easily migrate to Vite, right? So Rspack is perfect for that use case. obviously we believe for Greenfield projects we provide a overall a better set of trade-offs, but it’s good that you know people using Webpack still have options to modernize.

Kate Holterhoff (57:54)
Right.

Okay.

Evan You (58:09)
Yeah.

Kate Holterhoff (58:12)
Okay. All right. Fair enough. yeah, I will say when I was doing research on this, I could find most data from Vite, your team, and from Farm. And then there were a lot of like little like just individual developers who had who had put each through their paces, kind of created their own benchmarks. So you know, for as many numbers as there are, trying to like track down, I I found to be a bit of a challenge. So just just curious on what what you’re seeing. yeah.

Evan You (58:27)
Mm-hmm.

Yeah. Yeah.

I think that nowadays these benchmarks are more or less for tooling nerds. cause for end users, most of them, especially for nowadays if you’re just like vibe coding, like all these things are fast enough.

Kate Holterhoff (58:53)
Okay. Yeah, I I know. okay. I am gonna wrap us up here. I know we didn’t even broach AI nearly as as deeply as we could have, but we’ll just have to have you back on, Evan. okay, so before we do you direct folks to follow your musings? to what’s your social media channels? You know, how can everybody keep up with what’s new and exciting at VoidZero Cloudflare?

Evan You (59:19)
Sure, so I’m on X slash Twitter, I’m just @evanyou and that’s my main social channel, really. I have a I have a Chinese alt on X as well. That’s but it’s completely personal. It has nothing to do with technical stuff. But if you just want to follow for the kicks of it, you can do that.

Kate Holterhoff (59:42)
Yeah. We love kicks.

All right. Well, I have really enjoyed speaking with you today, Evan. Again, my name is Kate Holterhoff. I’m a senior analyst at RedMonk. If you enjoyed this conversation, please like, subscribe, and review the Monkast on your podcast platform of choice. If you’re watching us on Redmunk’s YouTube channel, please like, subscribe, and engage with us in the comments.