The Future of Image Formats: What to Actually Expect in 2026 and Beyond

Prediction articles about image formats have a poor track record. In 2019, multiple credible sources predicted WebP 2.0 would ship by 2021 and become the dominant format by 2023. Neither happened. In 2022, analysts predicted JPEG XL would reach full Chrome support by 2023. Chrome removed it in 2022 and only re-added it in 2025.

The incentives for writing confident predictions are obvious: they generate clicks and don't have expiration dates on them. The problem is that developers reading them make infrastructure decisions based on them.

This is a different kind of article. Here's what's actually happening in 2026, grounded in what's shipping in browsers, what's in specifications, and how the ecosystem is actually evolving—not how vendor roadmaps hoped it would.

---

The Settled Questions

Some things that were uncertain two years ago are now settled:

AVIF is the primary modern format for photographic web content. It has full support in Chrome, Firefox, and Safari, it's served reliably by every major CDN and image optimization service, and it consistently produces 40–60% smaller files than JPEG at perceptual equivalence. Developers starting new image pipelines today should use AVIF as the default for photographic content with WebP as a fallback.

WebP 2.0 is not coming. Google's open-source work on WebP 2.0 produced promising compression results in research contexts but was ultimately not pursued for standardization. The working assumption that "WebP will keep improving" was incorrect. WebP 1.0 is feature-frozen.

WebP is now the baseline, not the innovation. WebP does its job: it's universally supported, it beats JPEG, and it works without any fallback chain on modern browsers. But it's not where the performance gains are. It's the floor, not the ceiling.

---

The Active Transitions

JPEG XL is the most interesting ongoing transition. After being removed from Chrome in 2022 (citing insufficient adoption), it was re-added behind a flag and is now progressively rolling out to stable Chrome versions as of late 2025. Safari has had strong JXL support since iOS 17. Firefox support is growing.

What makes JPEG XL worth tracking is its portfolio of unique capabilities:

• Lossless JPEG transcoding: Convert any JPEG to JXL with 20–22% size reduction, with perfect reversibility. No other mainstream modern web format can do this.
• Progressive decoding: A genuinely well-engineered progressive loading implementation that improves perceived performance for large above-fold images.
• High-bitrate quality: JXL consistently outperforms AVIF at high quality settings, making it the better choice for photography-critical use cases.

The practical picture for JXL in 2026: support covers roughly 88–90% of browsers. Serving JXL with a <picture> fallback to AVIF is technically sound for early adopters. Teams operating at scale whose deployment or CDN pipeline doesn't include JXL encoding shouldn't feel behind.

Browser-native encoding is the less-discussed but more significant trend. As AVIF encoding via OffscreenCanvas.convertToBlob() has landed in Chrome and Safari, a new category of applications has become possible: client-side image optimization that produces AVIF output without any server or WASM overhead. This is what powers tools like TinyImage—the same encoding quality as server-side pipelines, running entirely in the browser tab.

This has format-agnostic implications: as new formats gain MIME-type support in browsers' Canvas APIs, browser-based tools can add them without infrastructure changes. The format pipeline improves at browser update cadence, not at service deployment cadence.

---

What Isn't Changing (Despite the Predictions)

JPEG will remain in use for 5+ more years. You can read breathless articles about "JPEG is dying" published every year since 2015. The format powers 60%+ of web imagery and has entire CDN networks optimized to serve it. Legacy content won't be retroactively converted. New content pipelines will gradually migrate to AVIF, but JPEG fallback chains will be necessary for a long time.

HEIC won't become a web format. Despite being the default format on iPhones, HEIC/HEIF faces patent licensing complexity that has prevented uncomplicated browser adoption. It works in the Apple ecosystem and as a device storage format, but it's not a realistic web delivery format.

AI-upscaling and ML-based compression are real but not transformative yet. Products like Google's JPEG-XL based ML compression pipeline (used internally) and Adobe's Firefly-based upscaling exist. They're improving. But they're not available to web developers as standard toolchain components. When they are, expect another 20–30% size reduction on top of current formats—but don't wait for them to start optimizing.

---

Building a Format-Agnostic Pipeline

The most important architectural insight from the last decade of format evolution: the specific format matters less than having an adaptive serving layer.

An adaptive serving layer means: you store your source images (ideally at high resolution RAW or TIFF quality), and your delivery layer generates the appropriate format at delivery time based on the Accept header.

Browser sends: Accept: image/avif,image/webp,image/jpeg,*/*

Server or CDN evaluates:
- If Accept contains image/avif → serve AVIF
- Else if Accept contains image/webp → serve WebP
- Else → serve JPEG

CDNs like Cloudflare, Fastly, and AWS CloudFront support this natively. Image optimization services like Cloudinary and imgix make it their baseline offering. Even the Vercel and Netlify image optimization layers handle this automatically for Next.js <Image> components.

The developer who builds a format-agnostic pipeline with this pattern in 2026 doesn't need to re-architect when JPEG XL reaches 95% browser support. They change one line of their format detection logic and the content starts flowing.

---

The Format Recommendation Matrix for 2026

Use Case	Primary	Fallback	Why
Photographic content, hero images	AVIF	WebP	Best compression at all quality levels
High-end photography, fine art	JXL	AVIF	Better high-bitrate fidelity
Legacy JPEG archive migration	JXL	AVIF	Lossless JPEG transcoding
Icons, logos, UI elements	SVG	PNG	Vector is format-agnostic
Animated content	WebP	GIF	AVIF animation is supported but complex
Screenshots, diagrams	WebP lossless	PNG	Better compression than PNG for these

---

What to Do Now

If you're managing an image pipeline in 2026, here are the actions ordered by actual impact:

1. Ensure your CDN serves AVIF to browsers that support it. This is the single highest-ROI change. If you're on a major CDN, it may already be enabled in settings.

2. Add width and height to every <img> tag. This prevents CLS regardless of which format you serve.

3. Preload your LCP image. One <link rel="preload"> tag in your document head. Works with any format.

4. Enable lazy loading for below-fold images. One attribute. Works everywhere.

5. Run a domain audit to find images that are being served in formats or at sizes worse than they should be. TinyImage's Website Optimizer does this across your entire accessible domain in under two minutes.

The format landscape will continue to evolve. The fundamentals of serving the right image at the right size in the most efficient format to browsers that support it—that won't change.