For most of my 14 years on bloated WordPress sites, “optimization” meant installing an image-smushing plugin and calling it a day. That stops working once you are dealing with zettabyte-scale data. The Future of Data Compression matters here because the standards we have leaned on since the 90s are hitting limits that no amount of caching will fix.
JPEG AI moves beyond pixels to latent tensors
JPEG has carried most of the images on the web for decades. The new JPEG AI standard (ISO/IEC 6048) works differently: instead of hand-crafted transforms, it uses end-to-end learning to map images into latent tensors. Smaller files are part of the point, but the bigger change is what you can do with a file without unpacking it first.
With traditional compression, you have to decode the pixels in full before a machine-learning model can look at them. That decode step is pure overhead on computer vision or object detection work. JPEG AI lets models operate directly in the compressed domain, so one stream can hand good-looking visuals to a person and raw feature data to a model at the same time, with no second decode.
MPEG VVC and the efficiency gap
On a high-traffic WooCommerce store with a lot of video, bandwidth is usually the biggest variable line on the bill. Versatile Video Coding (VVC/H.266) is the successor to HEVC, and it targets a 30% to 50% bitrate reduction at the same perceptual quality. That gap is what makes 4K and 8K delivery workable on the infrastructure you already have.
The compression ratio is only half the problem. The other half is energy per bit, and ISO/IEC 23001-11, better known as Green Metadata, finally addresses it. Devices can adapt their power draw based on what the content actually contains. Joules deserve about as much attention from us as TTFB gets.
Future-proofing your WordPress stack
Browser vendors will take their time with these standards, but your backend can be ready before they are. When mime types for JPEG AI or VVC land, WordPress media handling will reject the uploads by default. This snippet allows the new types in a custom build:
<?php
/**
* Prefix: bbioon_
* Logic: Allow next-gen compression formats in WordPress Media Library.
*/
add_filter( 'upload_mimes', 'bbioon_add_next_gen_mime_types' );
function bbioon_add_next_gen_mime_types( $mimes ) {
// Adding hypothetical JPEG AI and VVC extensions
$mimes['jpi'] = 'image/jpeg-ai';
$mimes['vvc'] = 'video/vvc';
return $mimes;
}
add_filter( 'wp_check_filetype_and_ext', 'bbioon_correct_next_gen_extension', 10, 4 );
function bbioon_correct_next_gen_extension( $data, $file, $filename, $mimes ) {
$ext = pathinfo( $filename, PATHINFO_EXTENSION );
if ( 'jpi' === $ext ) {
$data['type'] = 'image/jpeg-ai';
$data['ext'] = 'jpi';
}
return $data;
}
The final frontier: JPEG DNA
The strangest branch of the Future of Data Compression is biological. Data volumes keep climbing while magnetic and optical media last only decades. DNA holds information for centuries, which is why JPEG DNA is working out how to encode images as biochemical sequences robust enough to survive the errors that synthesis and sequencing introduce.
It sounds like science fiction, and it is also a practical answer to the “digital dark age” problem. For the near-term version of the same concern, I wrote up speculative loading, which handles data before it reaches the browser at all.
If keeping up with this is eating your dev hours, I can take it off your plate. I have been working with WordPress since the 4.x days.
The takeaway
Compression has stopped being a file-size trick. JPEG AI is aimed at machine vision, DNA storage is aimed at keeping archives readable a century from now, and both point away from human-centric pixels toward machine-centric semantics. The practical version for developers: stop treating “optimization” as a plugin you install, and start counting storage lifespan and energy cost in the same budget as page speed.
