Speculative loading is a background performance feature added to WordPress Core in version 6.8 that aims to make navigation feel faster by starting to load the next page before the browser normally would. Built and guided into core by Felix Arntz, the feature uses the browser Speculation Rules API to act on signals like clicks and hovers and to fetch resources during the brief delay between user intent and the browser’s default navigation fetch.
How it works
When someone clicks a link there is a short but measurable gap before the browser begins fetching the target page. Speculative loading uses that gap to begin retrieving the destination page so the user perceives a faster load. It relies on heuristics exposed by the Speculation Rules API and offers different levels of eagerness:
– Conservative (default in Core): Only starts speculative loads on click. This minimizes extra network activity while still gaining the millisecond-level advantage of beginning the fetch sooner.
– Moderate: Also considers hover signals on desktop and may start preloading a link when a user hovers before clicking. This can dramatically reduce perceived load time and, in some cases, produce an LCP of zero.
– Pre-fetch versus Pre-render: The API supports lighter pre-fetching, which fetches the HTML so rendering is faster when the user navigates, and heavier pre-rendering, which downloads and renders resources (HTML, CSS, JS, images) invisibly. Pre-rendering offers the fastest perceived navigation but can waste bandwidth if the link is not followed.
Browser support and progressive enhancement
The Speculation Rules API is currently implemented in Chromium-based browsers (Chrome, Edge, Opera). It is not yet available in Safari or Firefox. Speculative loading in WordPress is implemented as a progressive enhancement, so browsers that do not support the API are unaffected. The conservative, click-only default in Core reduces risk and unnecessary downloads across the diverse WordPress install base.
From plugin to core
Speculative loading first appeared as a plugin that collected real-world usage across tens of thousands of sites. That field data—along with telemetry and analysis—helped the WordPress performance team choose conservative defaults and validate expected gains. A plugin with a UI remains available for site owners who want more aggressive settings (for example, moderate eagerness with pre-rendering) or want a control panel to toggle behavior. The plugin can also revert sites to the safer core defaults.
Safety, exclusions and customization
WordPress Core deliberately provides no UI for speculative loading to keep a sensible default behavior for most sites and to follow a decisions-not-options approach. Developers, however, can modify behavior using hooks and filters. Sites can exclude specific URLs or patterns from speculative loading; exclusions may block pre-rendering while allowing pre-fetching where appropriate. Core also excludes URLs with query parameters by default because those are often associated with server-side actions that should not be triggered automatically.
Measuring impact with public datasets
The team used public, anonymized datasets to evaluate and quantify the feature. Two key sources are:
– Chrome User Experience Report (CrUX): Aggregated user performance data from opted-in Chrome users, accessible via BigQuery, which shows real-world experience metrics.
– HTTP Archive: Monthly crawls that capture site HTML and technology usage, useful for identifying WordPress sites and tracking configuration trends.
By combining CrUX and HTTP Archive data—isolating WordPress sites and tracking Core Web Vitals over time—the performance team estimated an approximate 1.9% improvement in relevant metrics for origins using speculation rules. That number is modest per site but meaningful at the scale of millions of WordPress-powered pages.
Trade-offs and environmental concerns
Speculative loading trades a small amount of extra network activity for improved perceived speed. Conservative defaults minimize wasted downloads, but site owners worried about bandwidth, privacy, or environmental impact can opt out, restrict pre-rendering, or apply speculative rules only to high-value links that analytics show users commonly follow. Agencies and enterprises often tune these settings based on site-specific traffic patterns.
Future and extensibility
Speculative loading is a foundation other developers can extend. Plugins, hosting providers, and agencies can build targeted rules using analytics and site knowledge to choose where speculation is most effective. As browsers add features and broaden support, more precise signals or device-aware heuristics may become available to further refine speculative behavior.
Conclusion
Speculative loading in WordPress 6.8 provides a mostly invisible but useful boost to navigation speed by leveraging browser signals to start loading the next page earlier. Conservative defaults make the feature safe for broad deployment, while an accompanying plugin offers more aggressive options and a UI for experimentation. Data-driven testing using CrUX and HTTP Archive validated the approach and helped quantify its impact. The end result is faster perceived page loads for users across many WordPress sites, achieved with cautious defaults and options for customization when needed.