WordPress market share sits at 42.2 percent of all websites in May 2026, down from a 43.6 percent peak in mid-2025. That is a 1.4 percentage point dip in twelve months, the first measurable softening since W3Techs began tracking in 2011. The honest read is that this is normalisation, not collapse. WordPress will be a core part of the web for the next two decades. The legacy is too deep to replace.
I co-run Seahawk Media. We have built and shipped well over twelve thousand WordPress sites since 2018, currently maintain around five thousand under care plans, and have been a WP Engine partner agency for most of that time. I also build modern-stack work, Astro, Next.js, Supabase, on the consultancy side. From inside both pipelines: Seahawk demand for WordPress maintenance and new WordPress builds in 2026 is consistent with 2024 and 2025. The dip in the W3Techs number is not showing up at our volume level. What is showing up is the future, and that future is AI-assisted everything. The most useful thing you can do this year is not pick a CMS. It is invest time and tokens in the AI coding tools that will compound for the next decade.
What the W3Techs numbers actually show
Three data points to anchor the conversation. W3Techs has WordPress at 42.2 percent of all websites with a known CMS as of May 2026, down from 43.6 percent in July 2025. HTTP Archive 2025 Web Almanac describes the same period as a shift from expansion to stabilisation, calling out that the WordPress installed base is now near-saturated. After fifteen years of growth, near-saturation is what a healthy plateau looks like, not the end of a category. The platform powers more sites today than it did at any point before 2024.
Astro on the other side: npm clocks Astro at roughly 2.5 million weekly downloads as of early May 2026, up from 1.4 million weekly downloads in May 2025. That is 100 percent year-on-year growth. Astro debuted in 2021 and crossed the one-million-weekly-downloads mark in late 2024. Next.js, SvelteKit, Nuxt and Remix all show similar curves at smaller absolute scale. These are real signals about where dev-led teams are spending build cycles, and they are worth taking seriously, but they are measuring developer interest rather than sites in the wild. The two curves do not invalidate each other. They describe two different parts of the market.
My prediction: a small further drop, two to three percentage points over the next two to three years, then a long stable plateau. WordPress finishes 2028 somewhere in the 39 to 41 percent band and stays there. That is not a dying platform. That is a category leader with a healthier competitive landscape around it than it has ever had.
Why the curve bent: three real drivers, not a death spiral
1. The market matured faster than the platform
The 2010s WordPress monopoly happened because the alternatives were either expensive enterprise CMSs (Sitecore, Adobe Experience Manager) or static sites that non-technical editors could not touch. That gap closed in the last three years. Sanity, Payload, Directus and Storyblok now offer modern admin interfaces that compete on editor experience while shipping a developer-friendly headless API. The competition is real, and that is a good thing for the open-source web. It also explains why the dev-led marketing-site category, the one place Astro plus a headless CMS now wins briefs that would have been WordPress in 2022, is the slice of the market where you see most of the W3Techs movement.
2. The Automattic vs WP Engine dispute is real but mostly already priced in
In October 2024 WP Engine sued Automattic and a federal judge granted a preliminary injunction in December 2024. The case remains active in May 2026. The dispute affected a lot of conversations in 2024 and the first half of 2025, particularly for buyers who had heard about it for the first time, and a portion of the W3Techs dip almost certainly reflects that period. By mid-2026 most of the buyers I talk to have either decided to stay with their existing WordPress agency, switched managed hosts, or moved on entirely. The story is no longer fresh news; the market is past peak panic. The platform is bigger than the dispute and the underlying open-source project is fundamentally robust.
3. AI-assisted coding rewrote the build math
This is the most important driver going forward, and the one to invest your time in. The AI coding tools that landed in 2024 and 2025, Claude Code, Codex, Cursor, Kimi, GitHub Copilot Workspace, are reshaping what a small team can ship. A dev-led team can build a forty-page marketing site on Astro plus Supabase in four to six days with serious AI assistance, where the same site might have taken three to four weeks of WordPress-builder work in 2023. That is a real productivity delta for a specific shape of project. It is also a delta that WordPress itself can capture. Gutenberg, the WordPress core editor, is being actively evolved. A WordPress core release that ships a modern, headless-first, AI-native admin surface would close most of the gap on Sanity, Payload and Directus inside one major version. I would not be surprised to see exactly that in 2027.
What I actually see in our Seahawk pipeline
Three honest observations from the agency floor.
First, demand for WordPress maintenance and new WordPress builds at Seahawk is consistent across 2024, 2025 and 2026. We are not seeing a drop in new WordPress briefs. The total volume of work is the same or slightly up, the queue depth is the same, and the customer mix is broadly the same. If WordPress were genuinely in trouble, our inbox would tell me before W3Techs would. It does not. The five thousand sites under care plans are growing in number, not shrinking, and the maintenance economics are the same as last year.
Second, the small-business end of the funnel, restaurants, dentists, family businesses, independent retailers, traditional service businesses, has not moved at all. The buyer there wants a CMS their non-technical office manager can edit, a familiar plugin set for forms and bookings, a WP Engine or Kinsta hosting bill they understand. WordPress is still the right answer for them, and that floor underneath the platform is wide and stable. This is the part of the market that protects the next two decades.
Third, the modern-stack work we win on the consultancy side is mostly net-new categories rather than WordPress migrations. Programmatic-SEO directory sites, AI-product company marketing sites, dev-tool company sites, content properties built around structured data at scale, that is where Astro and Next.js win briefs. Very few of those would have been WordPress projects in 2022. The two pipelines are running in parallel rather than competing for the same brief, and that is the shape of the market most people are missing when they look at the W3Techs chart in isolation.
Where WordPress will keep winning for the next two decades
Five categories where the WordPress moat is genuinely deep and will not be replaced quickly.
Plugin-economy use cases. Membership sites, learning management, donation processing, multi-vendor marketplaces, complex form workflows, appointment booking, restaurant menu management. The WordPress plugin economy is twenty years deep and there is no equivalent in the Astro or Next.js worlds. A small business that needs MemberPress plus WooCommerce plus Gravity Forms plus WPML plus a custom-fields integration is going to find that on WordPress in an afternoon. Building those features from scratch on a modern-stack base is a project, not a Saturday.
WooCommerce ecommerce at the small-to-mid scale. WooCommerce powers around 6.5 million stores globally, including a lot of small operators where the alternative is Shopify at thirty-plus dollars a month plus transaction fees. For a thousand-SKU store with thirty thousand orders a year, WooCommerce on managed hosting is still the right answer and will be in 2030.
Multi-author publishing with non-technical editors. National news rooms, magazine publishers, charity content desks. The Gutenberg block editor combined with role-based capability granularity is a genuinely good editorial surface that no headless competitor has fully matched at price. Sanity Studio, Payload Admin and Directus are close on the editor side, but none has the breadth of editorial workflow plugins, and the migration cost is high.
Small business CMS at scale. Local services, hospitality, retail, professional services. Easy to hire for, easy to maintain, easy to extend. The supply side of the agency market is mostly still WordPress-shaped, which means the maintenance, hosting and plugin economies have a long runway. This part of the platform alone is the floor that keeps WordPress at 38 to 41 percent of the web for the foreseeable future.
Education, NGOs, government and the public sector. Open-source, GPL-licensed, well understood by procurement, low licence cost, broad agency supply. WordPress wins these by default for reasons that have nothing to do with the technology curve and that are unlikely to change.
A WordPress renaissance is possible, and probably coming
The most interesting question for the next two years is not whether WordPress survives, it will, but whether WordPress reinvents itself again. The platform has done this before. The shift from classic editor to Gutenberg in 2018 was a generational rewrite that took two years of contributor effort and reshaped the entire builder ecosystem.
A second generational rewrite is on the table. A modern Gutenberg evolution that ships headless-first by default, a clean REST and GraphQL API surface that finally feels first-class rather than bolted-on, an admin UI rebuild that takes design cues from Linear, Sanity Studio, Notion and Directus, and an AI-native authoring experience that puts a Claude-class model behind every editor sidebar, would close most of the gap on the modern competitors inside a single major version. WordPress already has the editorial workflow, plugin economy and multi-role capability layer. Adding the developer experience and the AI-native authoring surface is real engineering work, not a research project. I expect to see most of that ship in WordPress 7 or WordPress 8 over 2027 and 2028, and at that point the W3Techs chart bottoms out and starts climbing again.
If you are betting on WordPress for the next decade, that bet is fine. If you are betting on WordPress staying exactly the way it is in 2026 for the next decade, that bet is probably wrong. The platform will move, and that is good news for everyone using it.
The actual move in 2026: invest your tokens, not your stack
The most useful thing you can do this year is not switch CMS. It is to spend serious time and tokens learning the AI coding tools that will compound for the rest of the decade. Five tools worth your attention, in order of where I would put my own time.
Claude Code is the agentic coding tool I use most. Long-running context windows, strong reasoning on multi-file refactors, native shell access. The 2025 and 2026 releases moved Claude Code from a curiosity into the most productive tool in my daily workflow. Worth ten hours of deliberate practice, then build the habit.
Codex is the GitHub-native counterpart, particularly strong on the GitHub Actions, pull-request review and merge-conflict workflow. If you live inside GitHub all day, Codex pays for itself in the first week.
Kimi from Moonshot AI is the rising long-context coding model and worth tracking. Strong on code synthesis at the file scale, particularly good with TypeScript, and the price per token is favourable for high-volume use.
Cursor is the editor wrapper that most of my dev friends defaulted to in 2024 and 2025. If you prefer the IDE-style workflow over the terminal-style workflow of Claude Code, Cursor is the right answer.
GitHub Copilot Workspace and the Vercel v0 family are the project-scaffolding tools that close the loop from rough idea to working repo. Useful for greenfield work, less useful for evolving an existing codebase.
The pattern across all of them is the same. Spend the first ten to twenty hours feeling slow. Build the habits. Learn which prompts and which files-in-context give you the right output. Once the workflow clicks, the productivity gain is real and it carries across whatever stack you are working in. WordPress, Astro, Next.js, Drupal, Shopify, the AI tooling is platform-neutral. The platform you build on matters far less than the workflow you build around it.
Should you migrate off WordPress in 2026? A balanced framework
Five questions to work through before you make the call. The right answer for most businesses is to stay on WordPress and modernise around it, not to migrate.
Question one: does your team have engineering ownership of the website? If yes, a modern-stack rebuild is on the table for the marketing surface. If marketing or content owns the site and engineering does not, stay on WordPress.
Question two: how many editors actually log in to publish? Three or fewer, all technical, modern stack is fine. Ten or more non-technical editors with role-based permissions, stay on WordPress and revisit when the WordPress 7 or 8 admin rewrite lands.
Question three: how complex is the runtime functionality? Content, forms, search, a few interactive elements, modern stack handles it. Membership, ecommerce, LMS, complex bookings, stay on WordPress for the long run.
Question four: how serious are the Core Web Vitals and INP requirements? If conversion rate from paid acquisition is measurably affected by load speed, the modern-stack delta is worth paying for. If organic SEO is the channel and the existing WordPress site passes Core Web Vitals, do not migrate; harden what you have.
Question five: is your team learning the AI tooling? If yes, that compounds across whatever stack you choose, and migration becomes a smaller decision than tooling discipline. If no, that is where your investment goes first, before anything platform-level.
For most businesses, the right play is hybrid: keep the WordPress backend, modernise the marketing surface on Astro or Next.js if the engineering case is strong, and put the saved time into AI workflow training across the team. That is the practical 2026 answer and it covers the next three years comfortably.
The honest two-decade view
WordPress is here for the next twenty years. The legacy, the plugin economy, the agency supply, the editorial workflows, the small-business floor, the public-sector adoption and the GPL licence depth combine into a moat that no current alternative breaches. A small dip in market share over the next two to three years is the price of a market maturing around a category leader. After that, my expectation is a long plateau in the 39 to 41 percent range, and a meaningful chance of a renaissance built around a modern admin and AI-native authoring that brings the chart back up.
For Seahawk, the answer remains what it has been since 2018: serve the WordPress buyer well, scale the modern-stack capability alongside, and bet on agencies that can credibly hold both. Demand for both is real. Demand for WordPress in particular has not moved at our volume level, and I do not expect it to in a meaningful way before WordPress 7 or WordPress 8 reshapes the platform from inside. The right move for the rest of 2026 is to invest your tokens, learn the AI tools deeply, and let the platform argument resolve itself over the next twenty-four months.
Common questions
Is WordPress dying in 2026?
No, not even close. WordPress still powers 42.2 percent of all websites and around 60 percent of all sites with a known CMS as of May 2026. The platform is in a measurable but small softening for the first time since 2011, down 1.4 percentage points from a 43.6 percent peak in mid-2025. That is normalisation around a maturing market, not a death spiral. My prediction is a further two to three percentage point drift over the next two to three years, then a long plateau in the 39 to 41 percent range. WordPress will be a core part of the web for the next two decades; the legacy and the plugin ecosystem are too deep to replace.
Is Astro really eroding WordPress users?
Astro is winning one specific slice of the WordPress base, dev-led marketing sites, documentation sites, programmatic-SEO directory sites, and content-led B2B properties where engineering owns the publishing surface. Astro is not a meaningful threat to plugin-economy WordPress use cases, WooCommerce ecommerce, multi-author publishing with non-technical editors, or the small-business CMS market. Astro's weekly npm downloads doubled to 2.5 million between May 2025 and May 2026; the slope is real, but the absolute footprint is still small. Treat Astro as a complementary tool for a specific kind of project, not a replacement for the WordPress platform.
What is the WP Engine vs Automattic lawsuit about?
WP Engine filed suit against Automattic in October 2024 alleging attempted extortion, abuse of power, and interference with business operations after a public dispute over WP Engine's contributions to WordPress core. A federal judge granted WP Engine a preliminary injunction in December 2024, restoring WP Engine's access to wordpress.org resources. The case remains active in May 2026. Most of the market impact was concentrated in late 2024 and early 2025; by mid-2026 buyer attention has moved on, and the underlying WordPress open-source project remains fundamentally robust. The platform is much larger than the dispute.
Should I migrate my WordPress site to Astro in 2026?
For most businesses, no. Stay on WordPress and modernise around it. The migration cases that make sense are dev-led marketing sites with three or fewer technical editors, no plugin-economy dependencies, and a paid-acquisition channel where Core Web Vitals materially affect conversion rate. Everyone else should keep the WordPress backend, optionally rebuild the marketing surface on Astro or Next.js as a hybrid, and put the saved budget into AI workflow training across the team. That is the higher-leverage move for 2026 and 2027.
Which AI coding tools should I invest in for 2026?
Five worth your time, in order. Claude Code for agentic multi-file work and long-running context. Codex for GitHub-native workflows including PR review and merge-conflict resolution. Kimi from Moonshot AI for long-context coding at favourable token economics. Cursor for IDE-style AI-assisted development. GitHub Copilot Workspace and Vercel v0 for project scaffolding from rough ideas. Spend the first ten to twenty hours on each one feeling slow, build the habits, then the productivity gain compounds across whatever stack you work in. The AI tooling is platform-neutral; the discipline is what matters.
Will WordPress survive the next twenty years?
Yes, with high confidence. Even at a 39 to 41 percent market share floor by 2028, WordPress would still be the single largest CMS on the web by a wide margin. The platform survives because of plugin-economy depth, WooCommerce breadth, editorial workflow maturity, and the inertia of thousands of WordPress-specialist agencies and millions of small businesses. The more interesting question is whether WordPress reinvents itself again over WordPress 7 and 8 with a modern, headless-first, AI-native admin surface that closes the gap on Sanity, Payload and Directus. I think it probably will, and at that point the W3Techs chart bottoms out and starts climbing again.