Is it true that you can ditch AMP to appear in Google Discover?

Was AMP really a requirement for Discover? \n\n

For a long time, confusion reigned on this point. Many SEO professionals believed — or feared — that AMP was a prerequisite for appearing in the Discover feed. This belief was based on empirical observations: AMP sites seemed overrepresented in this channel.

Mueller puts an end to this misconception. Discover accepts regular web pages, period. The myth of an algorithmic preference for AMP in this specific context was never officially founded. What actually mattered was the quality of the content, user engagement, and relevance to interests.

Why the confusion with thumbnails? \n\n

The misunderstanding stems from the thumbnail size. AMP pages benefited by default from larger thumbnails in Discover, which mechanically generated more clicks and visibility. This visual advantage was mistakenly interpreted as algorithmic favoritism.

However, Google has long offered the meta robots max-image-preview:large tag. It allows standard HTML pages to display exactly the same thumbnail sizes as AMP. In practical terms? A single line of code is enough to equalize display conditions. No need to overhaul an entire technical architecture.

What is the true hierarchy of criteria for Discover? \n\n

Discover operates on a model of interest-based personalization. The algorithm analyzes browsing history, past interactions, and contextual relevance signals. Loading speed matters, but it's just one factor among many — and AMP is merely one way to optimize this speed.

What truly matters: the quality of visual content, the freshness of information, measured engagement (click rates, time spent), and thematic coherence with user preferences. AMP can facilitate some of these criteria, but does not replace them. A fast, well-structured HTML page with quality images has every chance.

\n
• AMP is not a prerequisite to appear in Discover — standard HTML pages access it without restriction
\n
• The max-image-preview:large tag equalizes thumbnail sizes between AMP and non-AMP, neutralizing the historical visual advantage of AMP
\n
• The real selection criteria remain content relevance, user engagement, and quality of experience, not the technical format
\n
• Sites that invested in AMP solely for Discover may reconsider this strategy if maintenance becomes burdensome
\n
• Optimizing images, speed, and structured data remains more critical than the choice between AMP and standard HTML
\n

Is this statement consistent with real-world observations? \n\n

Yes and no. On paper, Mueller is right: technically, Google does not require AMP for Discover. But in practice, for years, AMP sites have dominated this channel overwhelmingly. Coincidence? Not really. AMP brought a set of optimizations — speed, optimized images, structured data — that maximized the chances of being picked.

Let's be honest: Google never stated that AMP was mandatory, but the ecosystem operated as if it were. The arrival of the max-image-preview tag changed the game, but many sites are still unaware of its existence. The result: AMP continues to benefit from a default advantage, even if it is no longer technically justified.

What nuances should be added to this assertion? \n\n

First point: saying that AMP is not required does not mean it is useless. If your tech stack is slow, your HTML is heavy, and your images are weighing in megabytes, AMP remains a shortcut to performance. It’s a constraining framework, but effective in enforcing good practices.

Second nuance: the max-image-preview tag only addresses a part of the problem — the thumbnail size. It does not compensate for a slow page, an overloaded DOM, or poor structured data. If your standard HTML is shaky, adding this meta tag won’t suffice to compete with a well-structured AMP. [To be checked]: Mueller does not specify whether other technical criteria still implicitly favor AMP in Discover's selection algorithm.

In what cases does this rule not apply? \n\n

If you are a media publisher with a massive content volume and a dominant mobile audience, AMP may still make sense — not specifically for Discover, but for your entire ecosystem (Google cache, guaranteed speed, integration with certain ad platforms).

On the other hand, for an e-commerce site, a corporate blog, or a niche site with low volume, maintaining two versions (AMP + HTML) quickly becomes a burden. Mueller's statement opens the door for simplification: focus on a single ultra-optimized HTML version, and you'll have the same chances in Discover. But beware — this assumes your HTML is truly top-notch. No compromises on speed or mobile-first.

What should I do concretely if my site uses AMP? \n\n

First step: evaluate the maintenance cost of AMP against the real benefit. If you find that Discover generates little traffic, or that your standard HTML is already fast (Core Web Vitals in the green), AMP becomes unnecessary. You can then consider a gradual migration to optimized standard HTML.

If you decide to abandon AMP, make sure your standard HTML incorporates all the technical signals that AMP brought automatically: structured data tags (JSON-LD), optimized images (WebP, lazy loading), loading speed under 2.5 seconds on mobile, and of course the famous meta max-image-preview:large. Without these foundations, your site risks losing visibility in Discover.

How can I check that my thumbnails display correctly in Discover? \n\n

Add the following tag in the head of your pages: <meta name="robots" content="max-image-preview:large">. That's it. No need to modify the rest of your code. Google Search Console does not provide a preview for Discover, so the only way to validate is to monitor your actual impressions in the GSC Discover report.

Also, check that your cover images comply with the recommended dimensions: at least 1200 pixels wide, 16:9 or 4:3 ratio, JPEG or WebP format. An image that is too small or poorly framed will be automatically dismissed, even with the tag in place. And remember to fill in the alt attributes and Open Graph tags — Google uses them to contextualize the image.

What mistakes should I avoid in this transition? \n\n

Classic mistake: suddenly removing AMP without preparing the replacement HTML. Result: traffic drop, disappearing thumbnails, degraded technical signals. Proceed in phases: first optimize your HTML, ensure performance is up to standard, then gradually remove AMP pages.

Another pitfall: believing that the max-image-preview tag is enough to compensate for a slow site. It only changes the thumbnail size, nothing more. If your Time to Interactive exceeds 5 seconds, your LCP is at 4 seconds, and your CLS is at 0.25, you will never pass Discover's selection filters, wide thumbnail or not.

\n
• Add meta robots max-image-preview:large in the head of all pages eligible for Discover
\n
• Ensure that the cover images are at least 1200px wide and maintain a consistent ratio (16:9 or 4:3)
\n
• Optimize Core Web Vitals (LCP < 2.5s, CLS < 0.1, FID < 100ms) to maximize selection chances
\n
• Integrate structured JSON-LD (Article, ImageObject) to help Google contextualize the content
\n
• Monitor the Discover report in Google Search Console to measure the real impact of changes
\n
• If migrating from AMP, maintain both versions in parallel for at least 3 months to compare performance before cutting definitively
\n