upvote
They put a massive amount of VC cash into convincing people that Next.js was "the modern way" to create a website. Then they got lucky with the timing of LLMs becoming popular while they were the hot thing, leading LLMs to default to it when creating new websites. To picture that amount of VC cash - they're at Series F, and a huge chunk of that went towards marketing.

Both have been changing as people realize it's rarely the right tool for the job, and as LLMs also become more intelligent and better at suggesting other, better options depending on what is asked for (especially Claude Opus).

reply
I really want this to be true. nextjs is a nightmare. I'm eternally disgruntled.

nextjs is also powerful due to AI. But the value is a robust interactive front-end, easily iterated, with maybe SSR backing, nothing specific to nextjs (it's routing semantics + React).

So much complexity has gone into SSR. I hate 5MB client runtime just to read text as much as anyone, but not if the tradeoff is isomorphic env with magic file first-line incantations.

reply
I have found SvelteKit really nice for SSR, and it avoids dealing with Vercel entirely.

Recent Claude models do well with it, especially after adding the official skill.

I have only recently started using it, so would love to hear about anyone else's experience.

reply
I don’t think they “got lucky”. nextjs is an old project now, and for a long time it was the simplest framework to run a React website.

This is why most open source landing pages used nextjs, and if most FOSS landing pages use it, then most LLM’s have been trained on it, which means LLM’s are more familiar with that framework and choose it

There must be a term for this kind of LLM driven adoption flywheel…

reply
> To picture that amount of VC cash - they're at Series F, and a huge chunk of that went towards marketing.

I guess they should have put some of that marketing money into hiring someone to manage the security of their systems. It's pretty telling that they had to hire an "incident response provider" just to figure out what happened and clean up after the hack. If you treat security like something you don't have to worry about until after you've been hacked you're probably going to get hacked.

reply
> they had to hire an "incident response provider" just to figure out what happened and clean up after the hack

Plenty to criticize them for, but that's totally standard and not something to ding them for. Probably something their cyber insurance has in their contract.

Forensics is its own set of skills, different from appsec and general blue team duties. You really want to make sure no backdoors got left in.

reply
> They put a massive amount of VC cash into convincing people that Next.js was "the modern way" to create a website

My impression is Next started becoming popular mostly as a reaction against create-react-app.

reply
So glad I decided to just stick with django/htmx on my project a few years ago. I invested a little time into nextjs and came to the conclusion that this can't be the way.
reply
You use a free template that's done in Next.js and uses its Image component, so you need a server.

Everything runs fine locally until you try to deploy it, and bam you need 4g ram machine to run the thing.

So you host it on Vercel for free cause it's easy!

Then you want to check for more than 30 seconds of analytics, and it's pay time.

reply
I am not following the logic. If you’re a hobbyist, sure.

But the argument is if you’re using Vercel for production, you’re paying 5-10x what you’d pay for a VM, with 4gb.

So then what’s the rationale? You can’t be a hobbyist but also “it’s pay time” for production?

reply
Perhaps the rationale is laziness. Maintaining VM probably takes some more effort and competence than deploying to Vercel. Some people are willing to pay to minimize effort and the need to learn anything.
reply
Vercel auto creates deployments on pushes to branches. That was a super useful feature in beta testing web stuff.
reply
Vercel promises to engineer the pain away when it comes to deployment. The thing however is that Vercel introduced that pain in the first place by writing sub-par documentation and splitting many of NextJS functions into small parts with different cost.
reply
Very nice developer experience. A lot of batteries included, like CDN, incremental page regeneration, image pipeline or observability. Not having to maintain a server.

I’m still planning to move elsewhere though, the vendor lock-in is not worth it and I’d like to keep our infra in the EU.

reply
All of this is available in Cloudflare $5 plan?
reply
Every three months I'm trying to deploy to Cloudflare from Monorepo and I hadn't have success yet. While Vercel works every time from the box. Maybe I could dig deeper and try to understand how it works, but I'm super lazy to do that.
reply
Cloudflare’s developer experience doesn’t come close, it is terrible. Cloudflare are working on it, and hopefully they’ll be a real competitor to Vercel on ease of use someday, but right now, it is painful when compared to Vercel. Cloudflare is infrastructure first, Vercel is developer experience first.
reply
Yes, CloudFlare's full of bugs and sharp edges. Not to mention the atrocious 3MB worker size limit (especially egregious in the age of ML models). They don't mention this up front in the docs and the moment you try to deploy anything non trivial it's oops time to completely re architect your app.
reply
> Not to mention the atrocious 3MB worker size limit

That's for the free plan.

Limits are documented here:

https://developers.cloudflare.com/workers/platform/limits/#w...

reply
Well it's so far from Vercel that it's not even funny any more.

Good work on workers though, maybe the next generation of sandstorm will be built on CloudFlare in a decade or so after all the bugs have been hammered out.

reply
In my experience it severely lacks on developer experience, compared to Vercel.
reply
For many people Vercel is Easy (not simple)

Knowing how to operate a basic server is perceived as hard and dangerous by many, especially the generation that didn’t have a chance to play with Linux for fun when growing up

reply
Great point on the playing with Linux growing up, it's second nature to me now.

I am always feeling like I'm doing something wrong running bare metal based on modern advice, but it's low latency, simple, and reliable.

Probably because I've been using linux since Slackware in the 90s so it's second nature. And now with the CLI-based coding tools, I have a co-sysadmin to help me keep things tidy and secure. It's great and I highly recommend more people try it.

reply
I haven't used Cloudflare and am the first to shit on Vercel. But I have to say, some aspects of their hosting are nice. In many ways it really is just a terminal command and up it goes with good tooling around it. For example, the PR previews take zero setup and just work. Managing your projects is easy, it's all nicely designed, it integrates well with Next and some other frontend-heavy systems and so on.
reply
If you are using nextjs it is easier because vercel done a lot of things to make it a pain to host outside of vercel.
reply
NextJs requires what exactly? Running a nodejs server? I mean yes, it takes a bit more time to set up than one-command deploy to Vercel. But in 2026, even this setup overhead can be cut down to minutes by telling your favorite LLM agent to SSH into your server and set it up for you.
reply
Do you have any examples?. I'm not that acquainted with the pains of deploying Next apps, though I've heard that argument being used.
reply
it's free for newbies and everyone, ofc it's a trap but freemium model gets people. aws can cost easily few thousands with 2-3 mistakes and clicks. vercel makes you start free then if you grow they bill you 10x-100x aws
reply
I dunno I put a lot of traffic through Vercel, maybe 100k visitors per day, and it was under a few hundred a month. I think a couple EC2 instances behind a load balancer would cost similar or more. I was under the impression that its still a VC subsidized service.

They regularly try to get me to join an enterprise plan but no service cutoff threats yet.

reply
For a lot of folks, I think its ease of deployment when using Next.js. I switched to astro, also doing a lot of cloudflare at the moment. Before that, I was doing OpenNext with sst.dev on AWS but it started feeling annoying.
reply
There really isn't any if you are running a serious product.

They have a free tier plan for non-commercial usage and a very very good UX for just deploying your website.

Many companies start using Vercel for the convenience and, as they grow, they continue paying for it because migrating to a cheaper provider is inconvenient.

reply
Out of curiosity what are you using cloudflare for that it costs $5 and who do you use for the baremetal box?
reply
I suppose their market is one click deployments. Maybe for non technical people or people not willing to deal with infra.
reply
Develop experience. Ephemeral deploys. Decent observability. Decent CI options. Generous free tier.
reply
Can one host a Next js app on cloudflare?
reply
Maybe. CF’s runtime isn’t perfectly identical to Vercel’s. For instance, CF doesn’t support eval(), which is something you shouldn’t be doing often anyway, but it did mean that we can’t use the NPM protobufs package that’s a dependency for some Google SDKs.
reply
I started using it a few years ago when I moved to my current company, and have to say I've learned to like it quite a bit. Moving to Cloudflare is an option, but currently it just works so we can't be bothered. Costs are not nothing, but basically no issues with it until now, and it's not so expensive that it raises eyebrows with the biggest being that we have 3 seats. The setup is quick and again it just works. We are a very small team, and the fact we don't have to deal with it on a daily/weekly basis is valuable. Obviously this current situation is a problem, but I am not sure which platform is free of issues like these. People act like it can't happen to me, until it does.
reply
It takes a while to realize you're being gaslit.
reply
0.82% of homes are burglarized every year.

Meaning since 2015, you’ve got an 8.2% chance of having someone walk out with that box. Hopefully there’s nothing precious on it.

reply
Assuming that all homes are at equal risk of being burglarized. In practice the neighborhoods I’ve seen are either at much higher risk or much lower risk.
reply
and burglarized homes have higher prob. of being burglarized again, and probabilities don't accumulate but compound, and is the server even in a house?
reply
They didn't imply the box was at their home and that probability is off
reply
If they have good backuos, no worries. Mine is in a locked colo cage in a datacenter, so I'm not worried either.
reply
I definitely do not keep it at home but the thought has crossed me for smaller less demanding boxes.
reply
That’s not how probabilities work.
reply
Imagining a thief walking in and demanding the home's RAM gave me a chuckle though.

Thieves probably look for small stuff like jewelry, cash, laptops, not some big old server.

reply
Or burglars.
reply
yes, this is indeed how probability works. thanks.
reply
>you’ve got an 8.2% chance of having someone walk out with that box.

The chance of being burglarized is not the same as the chance that when you are hit, they decide to take your webserver. Think it through.

reply