On Oct 23, 2026, Google’s John Mueller confirmed a simple but high-risk reality: JPEG XL (JXL) images currently won’t show in Google Images and won’t be used as preview images in Google Discover. His warning was blunt — “JPEG-XL is currently not supported
.”
If your site (or your CDN) starts outputting JXL as the primary format, you can accidentally switch off a chunk of image-driven visibility without touching titles, content, or backlinks. This is the kind of change that looks like “algorithm weirdness” in a dashboard and takes weeks to diagnose.
What happened (and what didn’t)
The story isn’t that JPEG XL is “bad.” It’s that Google’s surfaces that matter for most publishers and ecommerce brands — Google Images and Google Discover — aren’t consuming it yet.
Key detail business owners miss: this isn’t only about ranking in image search. Discover preview cards often pull an image. If Google can’t use the image you serve, your Discover presence can degrade or disappear even if the page is otherwise eligible.
Primary source: Search Engine Roundtable’s report on JPEG XL not being supported.
Why it matters (even if you don’t “do SEO”)
JPEG XL is getting pushed by well-meaning performance folks because it can be very efficient. The trap is assuming “smaller file” automatically equals “better SEO.” It doesn’t if the crawler or rendering pipeline can’t use the asset.
Here’s what we see in the wild: a dev team ships an “image optimisation upgrade” (often via a CDN toggle), the site gets faster in lab tests, and two weeks later the marketing lead asks why product images stopped appearing in Google Images — or why Discover traffic flatlined.
If you run ecommerce, publishers, recipes, news, sport, travel, or any brand that gets meaningful clicks from image-heavy SERPs, this is not a rounding error. It’s a distribution channel.
Our take: JPEG XL is overhyped for marketing sites (for now)
JPEG XL may be a great format in the long run. Today, it’s a compatibility bet. And marketing teams almost never get paid for “we saved 40KB per image” if it costs them visibility.
We usually recommend a boring, proven approach:
AVIF + WebP + JPEG/PNG fallback, served via <picture> with correct type attributes, and no surprises for bots.
Everyone will miss this: it’s not just your CMS. It’s your whole image supply chain — DAM, headless CMS, edge workers, CDN “automatic format,” WordPress plugins, Shopify apps, Next.js image components, and even your email team reuploading assets into a different system. Any one of those can start emitting JXL.
What this means for your site
- Check what you’re actually serving. In Chrome DevTools → Network → Img, confirm your production pages are serving AVIF/WebP/JPEG, not
image/jxl. Do this for category pages and top Discover-eligible articles. - Audit your CDN “auto” settings. If you use Cloudflare, Fastly, Akamai, or an image CDN, look for “automatic format” or “best format” features that might introduce JXL. Lock output formats explicitly.
- Keep fallbacks real, not theoretical. A fallback only helps if the bot can reach it. Test by fetching the HTML and ensuring the crawler sees a supported
src/srcsetoption. - Protect Discover-critical templates. For editorial templates that historically get Discover spikes, don’t experiment. Standardise image output and monitor changes like a release.
- Set a tripwire. Add a simple automated check in CI (or a weekly script) that flags
image/jxlresponses on key URLs.
Who wins, who loses
Winners: Teams staying on AVIF/WebP with solid fallbacks. Also, brands that treat “performance improvements” as controlled releases, not a toggle someone flips on a Friday.
Losers: Sites that outsource performance to a plugin/CDN and assume it’s safe. Also, orgs that measure SEO only by rankings — because this failure mode shows up first as missing thumbnails, lower CTR, and vanished image impressions.
What we’re changing on client sites this week
At BrandJunction, we’re treating this as a release-management issue, not an SEO theory debate. This week’s changes are practical:
| Risk area | What we check | What we change |
|---|---|---|
| CDN image negotiation | Response headers + MIME type for top templates | Disable/override JXL output; allow AVIF/WebP/JPEG only |
| HTML markup | <picture> order, type attributes, fallbacks |
Put AVIF/WebP first, then JPEG; ensure bot-visible img src |
| CMS/editor workflow | Upload paths (DAM → CMS → frontend) | Block JXL uploads or auto-convert to WebP/AVIF at ingest |
| Monitoring | Spot-check Discover + Images performance | Add alerts for sudden drops in image impressions/clicks |
If you need a team to own this end-to-end, it’s a mix of technical SEO and delivery discipline. That’s exactly the overlap our US build team for marketing and content sites spends most of its time on.
If you do nothing for 90 days
If you’re not serving JXL today, you might still get hit by an “automatic optimisation” update in your stack. The most common 90-day outcome isn’t a dramatic crash — it’s a slow, confusing bleed:
Discover stops picking up new URLs as often. Image thumbnails vanish on some queries. CTR drops on pages that used to win clicks visually. Someone blames “Google volatility,” and you lose a month chasing ghosts while competitors keep their previews.
And yes, this is happening in a period of broader search turbulence chatter. If you want the wider context from the same day’s coverage, see Search Engine Roundtable’s search news video recap.
Related context (worth 3 minutes)
Search Engine Journal also wrapped JPEG XL support into a broader SEO pulse (including Google’s notes on downtime status codes). If you’re coordinating dev maintenance windows, skim Search Engine Journal’s SEO Pulse covering JPEG XL plus 503 guidance.
Where BrandJunction fits (and what it costs to fix)
For most sites, fixing this is not a rebuild. It’s an audit and a controlled rollout: 1–3 days to identify where JXL is being introduced, then a small implementation to pin formats and add guardrails.
If you want a quick benchmark for engagement scope, our project and retainer pricing page outlines how we typically structure technical SEO + dev work without dragging it into a 12-week saga.
If you’re managing SEO across regions, we can also support implementation patterns consistently through our UK search optimisation work and our US-focused SEO delivery. We write about these “quiet breakages” regularly on the BrandJunction journal.
Next step: send us 5 URLs (home, a category page, and 3 top landing pages). We’ll tell you within 48 hours if you’re serving JPEG XL anywhere and what to change first.
Primary source: Search Engine Roundtable
