WordPress in 2026: why I still build on it (and the one thing that worries me)

Someone asked me recently, half-genuine and half-poking-the-bear: “Why are you still building on WordPress? It’s 2026. Can’t you just describe what you want to an AI and have it spit out a site?” Fair question. One of the better ones I’ve gotten lately, truth be told, and because the easy answer (“because I know it”) is also the laziest one, and I don’t want to be the guy defending his tool out of pure habit.

So I actually sat with it. Why WordPress, now, in the age of AI and the so-called idea person; the founder who can prompt their way to a working app over a weekend? It’s a little Jurassic Park out here: everybody’s so preoccupied with whether they can ship a thing from a chat window that nobody stops to ask whether they should be the one maintaining it six months later. What does a twenty-three-year-old PHP CMS offer that you couldn’t conjure out of an LLM?

Turns out, quite a bit. But let me get the uncomfortable part out of the way first.

Yeah, it’s losing share. And that’s fine.

WordPress is leaking market share, and it’s going to keep leaking it. As of this writing, it still runs about 41% of all websites on the internet, depending on what rag you’re reading, and roughly 59% of every site built on a known CMS, but that’s down from a peak a few years back, and the line’s pointed gently downhill.

I don’t think that’s the disaster the doomscrollers want it to be. The web grew bigger, and the tools improved. There are more good ways to build a site than ever before: Astro, Ghost, Framer, whatever AI builder is hot this month, and folks are picking the one that fits. Good! That’s healthy. The word of caution I have there is that walled gardens such as Framer, Webflow, and Loveable are going to fuck your shit up eventually; they’ll jack up your prices, reduce the value you receive somewhere (because they’re subsidizing your usage of their platforms with investment dollars. Be forewarned. Also, that’s the perfect case for owning your stack… WordPress shedding a few points off an absurd number isn’t a crisis; it’s not going anywhere… WordPress going up 1% would be a rounding error that any other platform would kill to call a great year.

So no, the share number doesn’t keep me up. The better question is the one underneath it.

So why still WordPress?

Strip away the habit and it comes down to a few things that never show up in a feature comparison.

It’s battle-tested. Boring word, huge deal. When I build something on WordPress, I’m standing on twenty-plus years of other people hitting the edge cases before me. The thing an LLM cheerfully generates in thirty seconds has hit exactly zero edge cases. It’s clever; it isn’t proven. There’s a real difference between code that runs and code that’s survived contact with the actual world, and I’ve been at this long enough to care a lot about the second one.

Then there’s the ecosystem and I want to be specific, because “WordPress has a lot of plugins” is the most boring possible version of this argument. I’m not talking about your favorite little single-purpose utility (love those and we build a bunch of our own). I’m talking about platform plugins. WooCommerce. Fluent. WS Form. The big, already-built engines that solve a genuinely hard problem so you don’t have to.

That’s the part the “just build it with AI” crowd waves away. Sure, an LLM can help me write a checkout. It cannot hand me 15 years of WooCommerce’s tax edge cases, payment gateway integrations, and refund flows that thousands of real stores have already stress-tested. Why would I rebuild that from a prompt? The wheel is right there, round and well-greased. The tools we have for building have gotten genuinely great, and the smart move is to use the good ones, not reinvent every single one because a chatbot made it look easy.

The part that actually keeps me excited

Here’s where I think people are sleeping on WordPress. There’s a new thing called the Abilities API, which shipped in WordPress 6.9 late last year. In plain English: it gives plugins a standard way to describe what they can do in a format an AI agent can read and act on. Pair it with the MCP adapter, and your favorite LLM, Claude, ChatGPT, whatever tool can access your site and actually do things through those plugins.

Sit with that for a second. The exact platform plugins I just praised for being built-out? They’re about to become things you can talk to. WooCommerce isn’t just for running your store; it’s becoming something an agent can query and operate on. The ecosystem doesn’t get less relevant in the age of AI. It gets more relevant because every one of those mature plugins becomes a capability your AI tools can plug into instead of rebuilding from scratch. That’s not WordPress surviving AI. That’s WordPress pulling up a chair and eating well.

The community is still the moat

One more, and it’s the squishy one nobody can put in a spreadsheet: the community. WordPress has a genuinely large, weird, loud, opinionated community that rallies around the thing at a scale I just don’t see anywhere else. And to be clear, I’m not dunking on anyone; Astro’s community is great, and Ghost’s is, too. Wherever you’ve got a tool with a vibrant community around it, you tend to do well. WordPress just happens to have an enormous one, and that’s a real, durable edge that won’t show up in any benchmark.

The one thing that worries me

Alright. I can’t write an honest piece about WordPress in 2026 and skip the elephant, so let me keep it tight: Matt Mullenweg.

I’m not going to relitigate the whole WP Engine saga — TechCrunch did a clean job of that, and you can go read it. But two things from it grind my gears real good, and I think they genuinely hurt the project.

One is the legal fight itself, a public, scorched-earth mess that did real damage to community trust, the kind you don’t win back with a tidy apology post.

Two — and this is the one that should bother anyone running a business on this thing, in the middle of all that, the quiet part got said out loud: WordPress.org is effectively Matt’s personal property. Not the nonprofit Foundation’s. His. Which means the place we all pull plugins and updates from, the thing we quietly treat as neutral public infrastructure, is one guy’s website that he can do whatever the hell he wants with, including deciding who gets kicked off when he’s in a mood.

How are you supposed to build a business on that with a straight face? You can — I do — but you’re accepting a real risk, and you should at least know you’re accepting it.

The fix is easy to describe and hard to make happen: .org needs to live under an actual independent board or nonprofit that Matt isn’t running, so it’s real public infrastructure instead of a personal playground. Until that happens, there’s an asterisk hanging over the whole thing, and pretending otherwise doesn’t make it go away.

Here to stay anyway

And yet — here’s the funny part — even with all that, the project keeps moving forward. It keeps shipping, keeps getting more capable, keeps innovating (see: the thing I got genuinely excited about three sections ago). The community drags it forward, sometimes despite its own founder.

So that’s my read on the state of WordPress: leaking a little share and totally fine about it, quietly better-positioned for the AI era than the hot takes suggest, with one big governance problem hanging over it that I really hope gets sorted out.

I’m still building on it. Not out of habit — out of a clear-eyed look at what it actually gives you. If you’ve got something complex to build, it’s still, for my money, the most sensible foundation going. Asterisk and all. Keep building.