Browsers do not wait for an entire web page to download before displaying it. Instead, they progressively process HTML, styles, and necessary JavaScript, then calculate layout, paint, and composite pixels. According to web.dev, the initial render typically requires some HTML, along with render-blocking CSS and JavaScript in the head. Browsers generally do not need to wait for the complete HTML, fonts, or images. The initial display does not mean the page is complete: it may still change as more resources load.
From HTML to the screen: what does the browser do?
The critical rendering path consists of the steps and resources required to complete the initial render. It does not cover everything needed to load the entire page. The rendering process can be summarized as follows:
- Build the DOM from HTML and the CSSOM from CSS, describing the document structure and styles, respectively.
- Execute JavaScript that changes the DOM or CSSOM, then use both to build the render tree.
- Calculate styles and layout to determine the positions and arrangement of elements.
- Paint pixels in memory, composite overlapping parts, and finally display the result on the screen.
This process does not happen just once. As resources that affect the display, such as images, load later, the browser repeats some or all of these steps to update what the user sees.
The initial render requires only a minimum set of resources
Browsers process HTML as a stream and can begin parsing as soon as part of it arrives, so they often start rendering before receiving the complete document. They need content to display, along with instructions for presenting it. The head provides metadata about the page and its resources, but browsers do not have to wait for every resource referenced there before displaying the page.
- Some HTML: Provides the document content to be rendered at that point.
- Render-blocking CSS in the head: Any necessary downloading and processing must be complete.
- Render-blocking JavaScript in the head: Only scripts that block rendering belong in this group; not every script is required for the initial render.
Displaying a page before necessary styles or scripts are ready can make it briefly appear broken. Waiting for every resource, however, can delay the appearance of a page that is already usable. Optimization should reduce unnecessary waiting, rather than remove all blocking resources.
Render-blocking does not mean the browser stops all work
CSS blocks rendering by default, but while waiting for it to download, the browser can continue processing HTML or doing other work. If a stylesheet’s media condition does not match the current environment, it need not block rendering. For example, when a page is viewed on a regular screen, the browser does not have to wait for print styles to load before displaying it.
Resources the browser need not wait for can still affect the experience
Browsers generally do not have to wait for fonts, images, or non-render-blocking scripts and styles outside the head before displaying a page. Scripts and styles outside the head must still meet the “non-render-blocking” condition; their location alone is not enough to determine this.
However, text may be temporarily hidden while fonts load, and image areas may remain blank. If sufficient space has not been reserved, particularly when the HTML does not specify image dimensions, resources that load later can cause layout shifts. So a resource can be important for displaying content even if it does not block the initial display.
If First Contentful Paint (FCP) or Largest Contentful Paint (LCP) is used as the measurement endpoint instead of first paint, the resources considered may also include those needed to display the relevant content, even if they do not block the initial render. The endpoint therefore affects which resources need to be considered.
How browsers render web pages: perspectives from across the web
Do render-blocking resources always block the entire page?
Not necessarily. web.dev notes that some browsers block only the content that follows a render-blocking resource. However, if such a resource is in the head, it effectively blocks rendering of the entire page. To assess the scope of its impact, consider not only whether the resource is blocking, but also its location in the document and how the browser handles it.
Why do some advocate using FCP or LCP as the measurement endpoint?
Because FCP and LCP focus on when content appears, rather than just the browser's first opportunity to display something. First paint may show only a background color, not meaningful content. Some therefore advocate using First Contentful Paint (FCP) or Largest Contentful Paint (LCP) and including resources needed to display the relevant content, even if they do not block rendering. First paint still has value as a measurement endpoint, however, because it captures the earliest opportunity for the user to see something on screen.
References
- Understand the critical pathweb.dev — Discusses: Do render-blocking resources always block the entire page?; Why do some advocate using FCP or LCP as the measurement endpoint?



