In short
A static site, such as one built with Astro, suits company websites whose content rarely changes. The server delivers prebuilt files without a database, which cuts response time and shrinks the attack surface. WordPress suits teams that publish often on their own or need ready-made features such as a store. What matters most is who handles content and updates.
Key takeaways
- In September 2026, 48% of WordPress sites worldwide passed all three Core Web Vitals on mobile, versus 71% for Astro and 78% for Hugo.
- Server response time drives most of the gap: only 23% of WordPress sites worldwide had a good Time to First Byte, versus 63% for Astro.
- Static doesn’t automatically mean fast: Gatsby sites transferred a median of 1.02 MB of JavaScript, and only 44% passed all three metrics.
- In 2025, a security vendor reported 11,334 new vulnerabilities in the WordPress ecosystem: 91% in plugins and six in the core.
- WordPress is the better choice when several people publish without a developer or you want ready-made features such as a store.
What’s the difference between a static site and WordPress?
A static site consists of prebuilt HTML files that the server delivers unchanged on every request. They’re built in advance, for example with Astro or another static site generator. WordPress, by contrast, uses PHP to assemble each page from templates and a database only when it’s requested. In return, you can change content directly in an editor.
A static site generator (SSG) is a program that turns text, written in a format such as Markdown, into finished HTML, CSS and JavaScript files. A site built this way has no server-side logic: everyone who requests a given URL gets the same content (MDN glossary). On a dynamic site, code on the server inserts data from a database into templates only when it’s needed (MDN, introduction to the server side). To do this, WordPress needs a server with PHP and a MariaDB or MySQL database (WordPress.org, requirements).
The term Jamstack for this approach originated at Netlify. Pages are prebuilt and served from a content delivery network (CDN), a group of servers spread across many locations (MDN glossary, CDN). Other features connect through APIs to backend services (Jamstack.org).
WordPress is by far the most widely used content management system (CMS). It’s used by 39.4% of all websites, or 58.1% of the websites whose CMS W3Techs can identify. Astro accounts for 0.2% of all websites and Hugo for 0.1% (W3Techs, Nov 8, 2026).
The line between the two isn’t clear-cut. With a caching plugin, WordPress can serve stored pages as static files (WordPress handbook, caching). Astro prebuilds pages by default but can also render individual pages on the server when they’re requested (Astro Docs, on-demand rendering). Our guide to technical SEO explains how search engines fetch pages and what else matters on the technical side.
| Criterion | Static site | WordPress | Sources |
|---|---|---|---|
| Load time for real visitors | all three Core Web Vitals good on mobile, worldwide: Astro 71%, Hugo 78%, Jekyll 78%, Gatsby 44% | 48%; the weak point is server response time | 8 |
| What the server does per request | delivers a prebuilt file | generates the page with PHP and a database, unless a cache serves a stored copy | 2, 3, 6 |
| Security | no login, database or application code on the web server; access credentials and the packages used in the build remain targets | security team for the core; in 2025, 11,334 new vulnerabilities, 91% of them in plugins, according to a security vendor | 16, 17, 18, 21 |
| Maintenance | nothing to update on the server; update the generator, Node.js and packages before the next build | core, plugins, themes and PHP on an ongoing basis; 15 releases in the latest series at the time, December 2025 to October 2026 | 22, 23, 24, 25, 26 |
| Cost over the site’s lifetime | software free; hosting for files; changes cost expert time unless a CMS is connected | software free; hosting with PHP and a database, licenses for paid extensions, ongoing maintenance | 3, 16, 27, 28; our assessment |
| Editing without a developer | only with a connected CMS (headless CMS); otherwise edit text files and rebuild | block editor, six roles, revision history | 1, 29, 30, 31, 34 |
| Extensibility | forms, search, login or an online store through external services or your own server functions | more than 78,000 free plugins in the official directory | 4, 7, 32 |
| Hosting | any web server, including via a CDN | recommended: a server with PHP 8.3+, MariaDB 10.11+ or MySQL 8.0+, and HTTPS | 1, 3 |
| Google and AI crawlers | content is already in the HTML | content is also in the HTML, which the server generates; many AI crawlers don’t see content that JavaScript loads later in the browser | 2, 37, 38 |
The numbers in the “Sources” column refer to the source list at the end of this article. Datafalke builds static websites, so this comparison relies on public data from third parties. Our own website appears only in the “From our work” section.
When is WordPress the better choice?
WordPress is the better choice when several people publish regularly without a developer. It’s also the better fit if you want to assemble features such as an online store from ready-made extensions, or if your team already knows WordPress well. In exchange, you have to plan for regular updates to the core, plugins, themes and PHP.
Strengths: The block editor has been the default since WordPress 5.0, released in December 2018. You use it to build pages and posts from blocks (WordPress documentation, block editor). Six predefined roles control who can write, publish or manage the site (WordPress documentation, roles). Every saved version stays in the revision history and can be restored (WordPress documentation, revisions). Many features come as extensions: the official directory lists more than 78,000 free plugins (WordPress plugin directory). According to W3Techs, 19.7% of WordPress sites use the WooCommerce online store extension (W3Techs, WordPress). On most sites, WordPress installs minor core updates automatically (WordPress handbook, updates).
Limits: According to the Web Almanac, a WordPress site’s speed and the effort it takes to run depend heavily on how it’s built. The report notes that page builders, which let you put together layouts without code, often produce more complex HTML and larger CSS and JavaScript files (Web Almanac 2025, CMS). Security vulnerabilities, too, are found mainly in plugins and themes (Patchstack, Feb 25, 2026).
Use case: blogs and magazines with several authors, clubs and associations with changing contributors, and online stores run by their own team. It also fits websites where the marketing team regularly creates new pages itself. If your team knows WordPress and an outside provider handles maintenance, we see little reason to switch.
When is a static site the better choice?
A static site is the better choice when your content rarely changes and an outside provider or a technically skilled person maintains it. It plays to its strengths when load time, a small attack surface and little ongoing maintenance are priorities. Typical uses include company websites, landing pages, blogs and documentation.
Strengths: Because the server doesn’t have to generate anything, it responds consistently fast. The prebuilt files can be distributed across several CDNs (web.dev, Rendering on the Web). With real visitors, these sites come out ahead. In September 2026, 71–78% of sites built with Astro, Hugo or Jekyll passed all three Core Web Vitals on mobile, compared with 48% of WordPress sites (HTTP Archive). There’s no login and no database on the web server, and there are no updates for you to install there.
Limits: Every change means editing a file, rebuilding the site and uploading it. You only get an editing interface if you connect a CMS, such as a headless CMS, which manages content but doesn’t serve any pages itself (Astro Docs, CMS). Search, login or payments come from external services. A static site calls them at build time or directly from the browser with JavaScript (Jamstack.org). And static doesn’t automatically mean fast: Gatsby sites, which load a lot of JavaScript, passed the Core Web Vitals less often than WordPress sites.
Use case: websites for service businesses, law firms, medical practices and trades businesses whose pages change a few times a year. Campaign landing pages and technical documentation are a good fit, too. Decide in advance whether an outside provider will handle changes or your team will use a connected CMS.
Is Astro better than WordPress for SEO and AI search?
Not automatically. Search engines and AI crawlers read content most reliably when it’s already in the HTML. Both WordPress and static site generators like Astro deliver that. Load time can help, because Google uses Core Web Vitals in its ranking systems. But Google says it still shows the most relevant content even when the page experience is weaker.
Both approaches deliver finished HTML: a static site builds it in advance, and WordPress generates it on the server. Problems start when content only appears once JavaScript runs in the browser. That’s why Google still calls server-side or pre-rendering “a great idea”: it makes websites faster, and not all bots can run JavaScript (Google Search Central on JavaScript). In a measurement that the cloud platform Vercel published in 2024, none of the major AI crawlers, including those of OpenAI, Anthropic and Perplexity, rendered JavaScript (Vercel, Dec 17, 2024).
Good Core Web Vitals, the three metrics for load speed, responsiveness and layout stability, don’t guarantee top rankings (Google Search Central on page experience). To check whether your content is in the HTML, see the section Can Google see content that JavaScript generates? in our technical SEO guide. Our article AI crawlers and robots.txt explains which AI crawlers exist and what blocking them does. To check this on your own website, use our how-to on JavaScript and AI crawlers.
Do static sites load faster than WordPress sites?
For real visitors, usually yes. In September 2026, 48% of WordPress sites worldwide passed all three Core Web Vitals on mobile. For sites built with the Astro, Hugo and Jekyll generators, the figure was 71–78%. But the generator alone doesn’t decide: only 44% of Gatsby sites passed.
The figures come from the HTTP Archive’s Core Web Vitals Technology Report. It combines measurements from real Chrome visits, taken from the Chrome UX Report (CrUX), with the technologies the HTTP Archive detects in its monthly crawl. A site counts as good if it meets all three thresholds at the 75th percentile of page loads. The thresholds are 2.5 seconds for LCP (load time of the largest element), 200 milliseconds for INP (response to input) and 0.1 for CLS (layout shift). If a site has no INP data, it’s assessed on LCP and CLS (HTTP Archive, September 2026 data). Our article on Core Web Vitals covers what each metric means in detail.
| Metric | WordPress | Astro | Hugo | Jekyll | Gatsby | All websites |
|---|---|---|---|---|---|---|
| Websites with data, worldwide | 2,658,175 | 47,277 | 6,043 | 1,894 | 9,278 | 8,569,446 |
| All three Core Web Vitals good, worldwide | 48.3% | 71.1% | 78.2% | 77.8% | 43.7% | 52.8% |
| Good LCP, worldwide | 55.4% | 81.1% | 87.6% | 87.3% | 64.4% | 65.6% |
| Good INP, worldwide | 90.3% | 85.0% | 89.6% | 85.7% | 76.0% | 79.4% |
| Good CLS, worldwide | 87.3% | 93.6% | 92.4% | 94.1% | 71.4% | 83.4% |
| Good server response time (TTFB), worldwide | 22.9% | 63.1% | 65.5% | 70.2% | 52.7% | 45.9% |
| All three Core Web Vitals good, visits from Germany | 63.1% | 82.7% | 89.6% | 89.2% | 54.3% | 69.2% |
| Websites with data, Germany | 172,438 | 3,276 | 490 | 93 | 676 | 586,975 |
Source: HTTP Archive, Core Web Vitals Technology Report (external site), · September 2026 data, mobile
For Germany, CrUX counts visits from users whose IP address is in Germany (CrUX documentation, countries). The figures therefore also include websites from other countries. For visits from Germany, every share is higher, but the order stays the same (HTTP Archive, Germany filter).
The share of WordPress sites with good scores is rising. A year earlier, in September 2025, only 45.0% of them passed all three metrics worldwide, or 59.9% for visits from Germany. Over the same year, though, Astro and Hugo sites improved more, from 61.6% to 71.1% and from 70.2% to 78.2%. Worldwide, the gap actually widened (HTTP Archive).
What causes the gap in load times?
Most of the gap opens up before the page even starts loading: worldwide, only 22.9% of WordPress sites had a good server response time, compared with 63.1% of Astro sites. A prebuilt file can be sent right away, while the server often has to generate a WordPress page first. On responsiveness to input, though, WordPress comes out ahead.
Response time: Time to First Byte (TTFB) measures the time from the request until the first byte of the response arrives. Up to 0.8 seconds counts as good. TTFB isn’t one of the Core Web Vitals, but it comes before the loading metrics First Contentful Paint (FCP) and LCP (web.dev on TTFB). According to web.dev, static rendering achieves a consistently fast TTFB because the server doesn’t have to generate the HTML first. With server-side rendering, that’s exactly the step that takes time (web.dev, Rendering on the Web).
Caching: The WordPress handbook calls caching “the fastest way to improve performance.” Caching plugins store posts and pages as static files and serve those instead. For pages that rarely change, this can improve performance “several hundred times over” (WordPress handbook, as of Oct 14, 2026). The HTTP Archive data doesn’t show how many WordPress sites use such a cache.
JavaScript: For responsiveness to input, 90.3% of WordPress sites had good scores, compared with 85.0% for Astro and 76.0% for Gatsby. The web.dev guide adds that prebuilt pages only respond quickly to input if they limit the JavaScript that runs in the browser. On mobile, Gatsby sites transferred a median of 1.02 MB of JavaScript and WordPress sites 0.68 MB. Astro sites came in at 0.41 MB, Hugo sites at 0.31 MB and Jekyll sites at 0.18 MB (HTTP Archive, page weight).
What the data doesn’t show: These are shares of groups of very different sizes: 2.7 million WordPress sites versus 47,277 Astro sites. The data doesn’t show whether switching would make a particular website faster. The HTTP Archive detects technologies automatically with a tool called Wappalyzer, and its URLs come from CrUX (HTTP Archive FAQ). The report doesn’t distinguish between Astro sites that prebuild their pages and those that render them on request. CrUX also includes only websites with enough Chrome visits (CrUX documentation).
Is a static site more secure than WordPress?
A static site has a smaller attack surface, because no login, database or application code runs on the web server. That doesn’t make it secure by default: access credentials and the packages used to build it remain targets. With WordPress, most vulnerabilities are in plugins and themes, and very few are in the core.
According to OWASP, an application’s attack surface covers all the paths that data or commands take into and out of it, plus the code that protects those paths (OWASP on attack surface analysis). A static site without a form has few of them: the server delivers files and accepts no input. WordPress has user accounts for editors and generates pages with PHP. Plugins and themes bring code of their own.
Figures from Patchstack, a company that sells protection for WordPress sites, show how this plays out. In 2025, 11,334 new vulnerabilities were found in the WordPress ecosystem, up 42% from 2024. Of these, 91% were in plugins and 9% in themes. Only six were in the core, and all of them were low priority. For 46%, the developer hadn’t released a fix by the time the vulnerability was publicly disclosed. The most heavily attacked vulnerabilities were first exploited a weighted median of five hours after they became known.
Source: Patchstack (external site), · vendor study
A security team of more than 50 experts looks after the core. Only the latest version is officially supported, but the team also backports fixes to older versions (WordPress.org on security). WordPress updates plugins and themes automatically only in special cases determined by its security team (WordPress handbook, updates). As of version 5.5, you can turn on automatic updates for each plugin and theme individually (auto-updates for plugins and themes).
A static site shifts the risks to access credentials and tools. According to the US Cybersecurity and Infrastructure Security Agency (CISA), a worm known as “Shai-Hulud” compromised more than 500 packages in the npm registry in September 2025. The malware searched for credentials such as GitHub tokens and keys for cloud services (CISA, Sep 23, 2025). Generators such as Astro get their packages from this registry (Astro Docs, installation). And as soon as a form accepts data, even a static site runs code on a server again.
How much maintenance does each option need?
WordPress needs ongoing maintenance on the server: the core, plugins, themes and PHP have to stay up to date, even when your content doesn’t change. A static site has nothing to update on the server. Instead, you update the generator and its packages, at the latest before your next change.
WordPress: From December 2025 to October 2026, WordPress shipped 15 releases of whichever series was newest at the time, including the major releases 6.9, 7.0 and 7.1. WordPress says only the most recent version in the 7.1 series is “safe to use and actively maintained” (WordPress.org, release archive).
On top of that, there’s PHP, the programming language WordPress is built on. On November 8, 2026, more than 30% of the installations that WordPress.org tracks were running PHP 8.1 or older (WordPress.org statistics). These versions no longer get security updates. For PHP 8.2, which another 22.6% run, security support ends on December 31, 2026 (php.net, supported versions). WordPress recommends PHP 8.3 or later (WordPress.org, requirements).
Extensions: For almost half of the vulnerabilities that became known in 2025, no update was available when they were disclosed (Patchstack, Feb 25, 2026). Back up your site before every update, as the WordPress handbook also advises, and check your most important pages afterward.
Static site: The server runs no site software that could go out of date. What you maintain instead is the toolchain that builds the pages. Astro, for example, requires Node.js 22.23.4 or later (Astro Docs). According to the Node.js project, long-term support (LTS) releases typically get fixes for critical bugs for a total of 30 months (Node.js releases page). If years pass between two changes, you’ll often have to move to new major versions before you can make the change.
What does each option cost over time?
The software itself is free in both cases: WordPress is licensed under the GPL, Astro under the MIT License. Costs come from hosting, maintenance, paid extensions and changes. With WordPress, they tend to be ongoing upkeep. With a static site, they tend to be the work involved in each change.
Calculate both options over the same period, such as five years, using these items:
- one-time cost of building the site or of a redesign
- monthly hosting × number of months
- monthly maintenance × number of months, including paid plugins and themes for WordPress
- changes per year × hours per change × hourly rate × number of years
- services for forms, search or a CMS, if your static site needs them
| Item | Static site | WordPress | Sources |
|---|---|---|---|
| Software | free, for example Astro under the MIT License | free under the GPLv2 or later | 27, 28 |
| Hosting | hosting or a CDN for files | hosting with PHP and a database | 1, 3 |
| Extensions | services for forms, search or a CMS, if needed | paid plugins and themes, if needed | 4, 16; our assessment |
| Maintenance | update the tools before a change | update the core, plugins, themes and PHP on an ongoing basis, make backups | 23, 25, 26 |
| Content changes | expert time, unless you use a CMS | your own time in the block editor | 29, 34; our assessment |
| Switching later | a new generator or CMS means a rebuild | older page builders make maintenance and switching harder, according to the Web Almanac | 15; our assessment |
Our cost guide lists market prices in euros for running a website in Germany, such as hosting and maintenance contracts, in its section What does a website cost to run?
Is there a middle ground between a static site and WordPress?
Yes. WordPress can serve prebuilt pages from a cache, or it can act purely as a CMS behind a static site. The other way around, you can connect a headless CMS to a static site so your team can edit text there. These hybrids combine advantages of both approaches, but they have more parts to maintain.
WordPress with a cache: A caching plugin stores pages as files and serves them without generating them again (WordPress handbook, as of Oct 14, 2026). You keep the editor and the plugins, and the maintenance, too.
WordPress as a backend CMS: Through its programming interface, the REST API, WordPress passes content to other applications (WordPress REST API Handbook). A generator can use that content to prebuild finished pages. Your visitors then get the prebuilt files, but WordPress still needs updates behind the scenes. Developers call this setup “headless.”
Static site with a headless CMS: The CMS manages text and images but doesn’t serve any pages itself; the generator fetches the content from it. Most of these systems offer a visual editor and a way to collaborate as a team (Astro Docs). The site is rebuilt after every change.
If you want a website without WordPress and without custom code, a website builder is another option. Our cost guide explains what a builder costs and who it’s enough for in its section Website builder or web designer: which costs less?
How do you choose between a static site and WordPress?
Decide based on how you’ll work day to day: who maintains the content, how often, and who handles updates? The seven questions below help you assess your situation. If most answers point to WordPress, it’s the better choice for you. If they point to a static site, you save ongoing maintenance and reduce your attack surface.
- Who changes the content? Several people without technical skills point to WordPress. An outside provider, or someone who’s comfortable working with text files, points to a static site.
- How often do you change things? Changes several times a week, for example for posts or promotions, point to WordPress. A few changes a year point to a static site.
- Which features do you need? An online store, a members’ area or a booking system with user accounts points to WordPress or custom software. A contact form and search work with either approach.
- Who installs updates? If no one is clearly responsible, that points to a static site. With a maintenance contract or a designated person, either works.
- How much do load time and attack surface matter? If they matter a lot, that points to a static site. With WordPress, a cache, a lean theme and few plugins help.
- What does your team already know? If your team knows WordPress well, you need a good reason to switch. Switching means a website migration, often with new URLs.
- What will it cost over five years? Calculate both options over the same period, including hosting, maintenance and changes.
If you plan to switch, set up a redirect for every old URL. Our website migration checklist covers what else matters. The questions and how we weigh them are our own assessment; no one has tested them as a whole.
From our work: what a static build means for datafalke.de
datafalke.de is a static site. Astro 7.4.4 builds the pages from Markdown files, and a CDN delivers them (technology in our colophon). In a check on November 9, 2026, none of the 98 pages listed in our sitemaps loaded JavaScript. The only script element on each page contained structured data in JSON-LD format. The Content Security Policy header our server sends with every page blocks scripts entirely: script-src 'none'. You can check this in your browser’s developer tools. How our contact form works without JavaScript and third-party services is covered in our how-to Contact form without tracking.
In the lab, we tested seven German pages from the site’s first version on November 7, 2026, shortly after launch, with one run each. Results vary from run to run, so our measurement log will report regular measurements as the median of five runs.
| Metric | Value across the seven pages |
|---|---|
| Lighthouse performance score | 99–100 |
| Lighthouse accessibility, best practices and SEO scores | 100 each |
| LCP in the lab | 1.1–1.7 s |
| CLS | 0–0.017 |
| Total Blocking Time | 0 ms |
| Requests per page | 7–8, all to datafalke.de |
| Data transferred per page | 99–133 KB, of which 69–87 KB fonts |
| Scripts loaded | 0 |
Measured · Lighthouse 12.6.1 in Chrome 156, mobile, simulated throttling, one run on each of seven pages
This way of building comes at a price. For us, every change, even fixing a typo, means editing a file, rebuilding the site and uploading it. We don’t have an editing interface. For the contact form, a small program on our host’s servers accepts each submission (see the colophon). We keep the build tools up to date ourselves.
There’s no field data for datafalke.de yet. The site has been online since November 7, 2026, and the Chrome UX Report only includes websites once they pass a number of visits that Google doesn’t disclose (CrUX methodology). We’re not claiming here that building this way brings us visits or inquiries. Only the measurement log will show that.
To check for yourself, test one of the German pages we measured, such as our home page at datafalke.de, in PageSpeed Insights or with Lighthouse in Chrome DevTools. Your timings will vary, because your device, network and software versions aren’t the same as ours. The number and size of the requests should stay close to ours as long as we leave the pages unchanged. Our guide to measuring page speed shows how to make your own Lighthouse runs comparable before and after a change.