All articles
MigrationJuly 7, 2026 6 min read

Moving Files Between Clouds Without Downloading Everything First

The download-and-reupload method breaks on slow connections, laptop sleep, and 100 GB folders. Here's how cloud-to-cloud transfer actually works.

Sooner or later everyone needs to move a pile of files from one cloud to another: leaving a full Google account, consolidating after a company switches from Dropbox to OneDrive, or moving an archive to cheaper storage like S3 or Backblaze.

The obvious method — download everything to your computer, then upload it to the new cloud — is also the worst one.

Why download-and-reupload fails

  • It doubles the transfer. 100 GB becomes 200 GB through your home connection. On a typical upload speed that's days, not hours.
  • It requires your machine to stay awake. Laptop sleeps, Wi-Fi hiccups, session expires — start over.
  • You need 100 GB of free disk you probably don't have.
  • Structure gets mangled. Zip downloads (Google Takeout, Drive folder downloads) flatten shared files, split into parts, and scatter metadata files everywhere.
  • No resume, no report. Which of the 40,000 files failed? You'll never know.

How cloud-to-cloud transfer works instead

A relay server with a fat connection reads from cloud A and writes to cloud B, streaming each file in chunks. Your computer just gives the order and can go to sleep. What to look for in any tool that does this:

  • Streaming, not storing — files should pass through the relay, never be kept on it. (CloudSync streams directly; 0 bytes are stored on our servers.)
  • Retry and resume — transient API errors are normal at scale; the tool should retry failed files automatically and resume interrupted runs instead of restarting.
  • A per-file report — copied, skipped, failed, with reasons.
  • Incremental re-runs — the second run should only move what changed, so you can migrate in waves and cut over cleanly.

A realistic migration playbook

  1. Inventory: connect both clouds and browse them side by side (CloudSync's Unified Explorer does this) to decide what's worth moving.
  2. First wave: create a one-way incremental sync from old → new and let it run overnight. Big folders continue across runs automatically.
  3. Delta wave: re-run before the cutover — only files changed since the first wave get copied.
  4. Verify: check the job report for failures, retry them, spot-check a few folders.
  5. Keep the bridge: instead of deleting the old account immediately, keep a scheduled sync for a month as a safety net.

CloudSync's free plan is enough to run a small migration end-to-end and see the full flow before you commit.

Put this into practice

Syncrix connects 13+ clouds and keeps them in sync automatically. Free plan, no credit card.