WordPress has lost market share for six months in a row. AI website builders like Lovable, Framer, and Cursor can spin up a site from a sentence. The WordPress CEO spent last year publicly fighting the host that powers a big chunk of the ecosystem. So the question is fair: is WordPress dying?
The short, honest answer: no. WordPress is not dying. It is shrinking slightly, it is being reinvented by AI from inside, and the people who bet their careers on it are going to be fine — if they understand what is actually moving. This piece is the version I wish I had read before I started worrying.
The headline number nobody reads correctly
WordPress powers 41.2% of all websites as of 29 July 2026, and 59.1% of all websites that use a known content management system, according to W3Techs. That share has fallen from 43.20% in December 2025 to 41.90% by May 2026 — six consecutive months of decline, and the pace is accelerating (more than double the previous year's drop, per Search Engine Journal).
That sounds bad until you line it up against the competition:
| Platform | Share of all websites | What it gained in the same 6 months |
|---|---|---|
| WordPress | 41.2% | −1.3 pts |
| Shopify | 5.2% | +0.2 pts |
| Wix | 4.3% | +0.1 pts |
| Squarespace | 2.5% | +0.1 pts |
| Joomla | 1.2% | (~flat) |
WordPress is still roughly eight times larger than its nearest competitor. Shopify, Wix, and Squarespace grew — but together they added less than half a point. The share WordPress "lost" did not migrate to a rival CMS. It went to a different category entirely: AI-built and static-generated sites that leave no CMS fingerprint at all. That is the part that should make you think.
Why the numbers went sideways: four real reasons
The decline is not random noise. There are four identifiable drivers:
- AI website builders got good enough for simple sites. Lovable (formerly GPT Engineer) ships deployable React apps from a chat prompt; Framer does the same for design-led marketing sites. If you need a single landing page, you no longer need a CMS at all.
- The Mullenweg-WP Engine feud damaged confidence. In late 2024, Automattic CEO Matt Mullenweg launched public attacks on WP Engine, blocked their customers from plugin updates, and required contributors to disavow ties to the company. The fallout hit the next quarter's market share data directly.
- WordPress has maintenance tax. Core updates, plugin conflicts, 11,334 new vulnerabilities disclosed in 2025 (91% from plugins, per Patchstack). Small businesses increasingly want to "manage a business, not a website," as one Shopify developer put it.
- Developers are sampling new stacks. Astro downloads more than doubled over the same period — from 4.59M in January to 9.24M in April. Some developers are leaving WordPress not for Shopify, but for newer developer-focused tools.
Notice what is not on that list: any sign that WordPress itself is technically broken. The platform is fine. The story is the world around it moving slightly faster than the platform's growth.
The most important thing nobody is talking about: AI came INTO WordPress in May 2026
Here is the part that changes the whole conversation. While everyone was debating whether AI would kill WordPress, WordPress shipped AI into its own core.
WordPress 7.0 "Armstrong" (yes, named after Louis Armstrong the jazz musician) was released on 20 May 2026. It was not a quiet release: 875+ contributors, 420+ enhancements, the first major version of the 7.x series. The headline from the release announcement was explicit — "WordPress 7.0 marks the start of a new era, laying the foundation for AI across the WordPress experience."
What that means in concrete terms:
- A new AI Client built into Core lets plugins, themes, and Core features talk to generative AI models without bundling their own SDKs.
- The Abilities API exposes a server-side interface so any plugin can request an AI capability (summarize, draft, transform, classify).
- Native connectors for OpenAI, Google Gemini, and Anthropic Claude ship out of the box.
- Version 7 is now the dominant release — adopted by 52.8% of WordPress sites, with Version 6 holding 39.9%. People are actually upgrading.
Real-time collaborative editing was pulled late and is expected in WordPress 7.1, August 2026, but the AI foundation is already live. This is the opposite of "dying." This is the platform absorbing the technology that was supposed to replace it.
The AI + WordPress pipeline that actually works in 2026
This is where the "AI kills WordPress" framing completely collapses. In practice, the AI-to-WordPress pipeline is a workflow, not a war. It looks like this in 2026:
- Generate the design with an AI agent. Claude, Cursor, or Gemini writes a design prompt, builds the components, and produces a layout no human could draft in anywhere near the time.
- Export the build into WordPress with NovaMira. NovaMira is a free (with a €49/year Pro tier ~$53 USD) WordPress plugin that gives AI agents full access to WordPress — Claude, Codex, Cursor, Gemini CLI, Antigravity, even ChatGPT. It builds inside Elementor, Bricks, Divi, Oxygen, Beaver Builder, and the block editor — not by suggesting code, but by actually opening your page builder and writing into it.
- Refine with a real WordPress page builder. Elementor still powers 31.5% of WordPress sites according to W3Techs — and now has an AI assistant called Angie that builds sites and custom widgets from a prompt (we covered it here).
- Extend with the plugin ecosystem. Yoast SEO (13M+ active installs), Rank Math (4M+), WooCommerce (19.8% of all WordPress sites, ~5M+ active installs), WPBakery, and 64,000+ free plugins in the official directory. No AI website builder has this.
The AI does the slow, expensive part (design and code generation). WordPress does the part AI is bad at (security, plugin ecosystem, SEO tooling, hosting portability, e-commerce, long-tail extensibility). The two are partners, and a designer who can run this pipeline can ship a client site in a fraction of the time it took in 2024.
If you want a head-to-head of where each AI website builder sits — Lovable, Framer, Bolt.new, v0, and the WordPress + NovaMira pipeline — I wrote that comparison here: Best AI Website Builders 2026.
The worst-case scenario — and why it is not scary
Let us take the doom argument seriously. Suppose WordPress "pulls the plug" in some hypothetical future. What actually happens to your self-hosted WordPress site?
Almost nothing, on day one. WordPress is open-source software, downloadable from wordpress.org, running on your own hosting. There is no kill switch at Automattic that can shut off your site. The PHP files on your server keep executing, the database keeps serving pages.
The real risk shows up later, in months and years, if updates stop coming. Untimely plugin updates and unpatched security holes compound. But that scenario is also where the migration opportunity lives.
If WordPress's share really did decline further, the WordPress-to-something-else migration market becomes huge overnight. There are roughly 475 million WordPress sites in the world (per various trackers citing W3Techs). Even if only 5% of them ever migrated, that is 23 million sites needing conversion tools, SEO preservation, content export, 301 maps, and theme rebuilds. That is a business opportunity — not an apocalypse. The plugin ecosystem itself is worth billions; the managed WordPress hosting market sits at $8.74 billion in 2025, growing toward $13B by 2032.
So: the downside of being a WordPress professional is not unemployment. It is redirecting one service line.
The lesson that mattered most to me: attach to the problem, not the tool
A few years ago, my own YouTube channel hit a wall. I had built a career on long, structured WordPress tutorials — 14-minute setups, click-by-click. When AI changed how people learn, those tutorials did not perform the same way. I had two choices: try harder at the same thing, or adjust. I adjusted. I moved toward what some people call "vibe coding" — showing how AI agents connect to WordPress and what actually works in 2026, faster. I still teach the same skill (building an online presence for your business); the format changed.
Sometimes it is better to adjust instead of try harder. That is themeta-lesson for web professionals in 2026. Tools change. The client's problem does not. SEO does not change. Design sense does not change. Knowing how to scope a project, manage a client, price a deliverable, preserve rankings across a migration — none of that depends on WordPress. The skills transfer; the tool may not.
What I would actually do in 2026
If you are an existing WordPress professional:
- Upgrade to WordPress 7.0 on staging, verify your plugin stack, then roll out. Native AI infrastructure is now table stakes.
- Wire NovaMira to your existing AI client (Claude, Cursor, or Codex). The plugin is free; Pro is €49/yr for Elementor and Bricks support. The competitive advantage is not the AI — it is the AI-plus-WordPress pipeline.
- Get disciplined about security. 11,334 vulnerabilities disclosed in 2025 and 91% are from plugins. Updates are not optional; they are the job.
- Price for outcomes, not tooling. Migration, SEO preservation, and AI-augmented redesigns are the premium services in 2026, not "I built your WordPress site."
If you are a small business owner deciding what to build on:
- For a marketing site or a blog that needs real SEO control: WordPress + a modern page builder (Elementor, Bricks) + an AI client (Claude, Cursor) + NovaMira is still the most flexible and the most future-proof.
- For an e-commerce store: consider Shopify. WooCommerce powers 5M+ stores and stays strong, but Shopify is purpose-built and outscores WordPress on simplicity for a single-shop operator.
- For a single landing page or a SaaS prototype: use Lovable, Framer, or Bolt. Don't fight WordPress for the simplest possible case.
The verdict
WordPress is not dying. It is in the middle of the most significant reinvention in its 22-year history — absorbing AI into core, modernizing the dashboard, and repositioning a vast plugin ecosystem toward AI-native workflows. Market share is down a little. The platform is up a lot.
The honest answer to "is WordPress dying" is: no, but the WordPress of 2024 is gone. The WordPress of 2026 speaks to AI agents, ships native connectors for OpenAI, Gemini, and Claude, and competes with the new builders by integrating with them instead of ignoring them. If you have been waiting for a sign to learn the new pipeline, this is the sign.
This article was adapted from the YouTube video "Is WordPress Dying? Here's What's Actually Happening" by Ferdy Korpershoek and expanded with 2026 W3Techs data, the WordPress 7.0 release notes, and verified market figures.
Discussion
0 comments