A look at the software and methods used to distribute 11M documents, totaling 2.6TB, of the Panama Papers to 100+ publications and about 400 journalists
From Encrypted Drives To Amazon's Cloud — The Amazing Flight Of The Panama Papers — It was an epic haul.
Context & Ripple Effects
This Forbes explainer lands one day after Australia's Nuix donated its document-analysis software to ICIJ, completing the picture of the Panama Papers pipeline: encrypted drives and Amazon's cloud moved 2.6TB across borders, while donated tooling made 11M documents searchable by roughly 400 journalists at once.
The significance is methodological as much as journalistic — this was an early demonstration that a single mega-leak could be processed collaboratively by 100+ publications using commodity cloud infrastructure rather than one newsroom's secure room.
First-order effects
- ICIJ's network of about 400 journalists and 100+ publications gains simultaneous access to the full 2.6TB corpus over Amazon's cloud, replacing physical media handoffs with centralized, searchable distribution.
- Amazon's cloud quietly becomes critical leak-handling infrastructure for investigative journalism, alongside the encrypted-drive logistics described in the piece.
Second-order effects
- The distributed-corpus model becomes a template for transparency groups: Distributed Denial of Secrets later publishes 175GB of leaked Russian documents through a WikiLeaks-style repository, extending bulk-leak distribution beyond curated press coalitions to public dumps.
- Bulk data distribution scales well past journalism's needs — Anna's Archive's release of ~300TB of Spotify-scraped torrents shows the same move-large-corporpora-publicly playbook now operating outside any editorial gatekeeping.
Third-order effects
- As leaks are staged on shared storage, the security perimeter shifts from couriers to configuration: ZDNet's finding of ~2.3B files exposed on publicly accessible storage servers shows the failure mode of cloud-staged sensitive data at scale.
- If the pattern holds, large-leak handling splits into two tracks — vetted journalist coalitions on commercial clouds, and open repositories publishing raw dumps directly — with the line between them drawn by whoever controls access to the corpus.
The trend: Mega-leaks are shifting from couriered hard drives to cloud-distributed, searchable corpora shared across hundreds of collaborators — with uncurated repositories now scaling the same approach to the general public.