How macOS actually works

When a Mac has too many files

Updated 2026-08-05 · 6 min read

Most search advice assumes a few hundred thousand files. Past a few million, different things break.

Who has this problem

Photographers, designers, video editors, developers, anyone with a decade of archives. One customer, a solo design studio doing web work, print and retouching, had roughly 62 million files across six drives. That is not exotic in creative work — a single shoot produces thousands of files, and nothing is ever deleted.

What breaks first

Spotlight quietly gives up. Not with an error — it simply stops returning things, particularly on external volumes.

Scanning tools become unusable. EasyFind and Find Any File read the disk on every search. Fine at 200,000 files; unusable at 6 million, and hopeless at 60 million.

Naive indexers run out of memory. An index needs roughly 45 bytes per file, so 62 million files is about 2.8 GB before overheads. That is real memory pressure on a 16 GB Mac, and swapping makes everything slower still.

What actually works at this scale

A compact index and a memory-mapped read. At 45 bytes per record, six million files is a 277 MB file — mapped rather than loaded, so the kernel pages in what is touched.

An index that skips work. Scanning six million records per keystroke is roughly 28 ms: too slow to feel instant. A trigram index that eliminates most of the file before scanning brings a distinctive query to about 1.4 ms. That difference is the entire experience.

Excluding what is duplicated. Time Machine volumes and backup clones are copies of files already indexed. Excluding them is the single largest reduction available, and on the 62-million-file Mac above it was most of the total.

What to do if this is you

1. Exclude Time Machine drives. Always, and first.
2. Decide about backup clones. If MacHD-Backup duplicates your boot disk, indexing it doubles your index to tell you things twice.
3. Index archive drives once, then unplug them. A tool that keeps unmounted volume records tells you which drive a file is on without connecting it.
4. Search with two terms. At this scale one common word matches tens of thousands of things.

The uncomfortable truth about counts

A file count is not a measure of your work. Node modules, caches, browser profiles, app support folders and photo library sidecars produce millions of files nobody will ever search for. Most large Macs are large because of machine-generated files, not because their owner made sixty million things.

Which is why exclusions matter more than raw index speed. The fastest search is the one that never had to consider the file.

Common questions

How many files can a Mac search index handle?

A compact index costs roughly 45 bytes per file, so six million files is about 277 MB and sixty million about 2.8 GB — real memory pressure on smaller Macs.

Why is my Mac's file count so high?

Usually machine-generated files: node_modules, caches, browser profiles, app support folders and photo library sidecars, not documents you created.

What is the fastest way to search millions of files?

An index that can skip most of itself. Scanning six million records per keystroke is around 28 ms; a trigram index brings a distinctive query to a few milliseconds.