File formats look frozen from the outside — the same PDFs, JPGs and ZIPs as a decade ago. Under the surface, the practical landscape shifts constantly: which formats browsers support, which tools default to, which archives accept. Here is where things genuinely stand, and what is likely to change for people who work with documents and images every day.
Images: the WebP decade is real
The clearest winner of the last few years is WebP. With universal browser support since 2020-era Safari releases, it now does what JPG and PNG each do separately — lossy and lossless modes, full alpha transparency — at 25–35% smaller sizes than JPG at comparable quality. The practical effect is already visible: major sites serve WebP first with JPG as fallback, and tools increasingly default to it. The older formats are not disappearing — decades of cameras, scanners and software produce JPG and PNG forever — but "convert to WebP" is steadily becoming the default optimisation for anything destined for the web.
Documents: open standards keep winning slowly
In documents, the long-running story is open standards displacing vendor-locked ones, slowly and unevenly. DOCX won its format war years ago and is readable everywhere; ODT survives in open-source ecosystems and government mandates. The genuinely important open standard is PDF itself — ISO 32000 — and especially its archival profile PDF/A, which more archives, courts and agencies now require for long-term storage because it embeds fonts and forbids external dependencies. Expect PDF/A requests to keep growing anywhere documents must outlive the software that made them.
The cloud changed the default, not the file
Google Docs and Office 365 moved collaboration into the cloud, and their native formats are cloud documents, not files. Yet exports still end in the same places: DOCX for exchange, PDF for anything final. The lesson from a decade of "the file is dead" predictions is that files turn out to be the durable boundary between systems — the format you hand to someone whose tools you don't control. Conversion between those boundary formats is, if anything, more central now than before.
What is NOT happening (honest predictions)
- PDF is not going away. Too much law, logistics and printing depends on it. Its tooling keeps improving; its role doesn't change.
- Legacy formats die slowly. DOC (pre-2007), RTF and Lotus formats still arrive in inboxes; support for reading them outlives support for creating them by decades.
- AI changes conversion tooling, not the formats. Better OCR, smarter layout reconstruction and auto-tagging improve results — the target formats stay the same. Where machine learning genuinely helps today is scanned-document recognition and table reconstruction.
What to actually do about it
Three practical habits cover the trends. Prefer open, widely-supported formats for anything you create — DOCX over DOC, PDF/A for archives, WebP for web images. Keep masters in lossless or editable forms (the PNG before the JPG, the DOCX before the PDF). And when you receive something odd, convert it to something boring as early as possible: every conversion problem gets cheaper the sooner the file is in a standard format. When you need one of those conversions, the converter handles the formats in this article — and the boring ones will keep working for years.