A brand new one terabyte drive comes out of the box, gets plugged in, and the computer cheerfully reports something closer to 931 gigabytes. It looks like sixty-nine gigabytes vanished before a single file was saved. Nothing is broken, and nobody took the space; two different counting systems and a few unavoidable reservations explain the whole gap. Once the arithmetic is clear, the shrinkage stops looking like a defect and starts looking predictable, which helps when you are sizing a drive for a real workload rather than trusting the number printed on the box.
Two different ways of counting a byte
The gap starts with a units disagreement that goes back decades. Drive makers count in decimal, where one kilobyte is 1,000 bytes, one megabyte is 1,000 kilobytes, and one terabyte is exactly 1,000,000,000,000 bytes. Computers, by contrast, group memory in powers of two, where the tidy number is 1,024. So the operating system divides that same trillion bytes by 1,024 three times and lands on about 931 of its own units.
To reduce the confusion, standards bodies introduced separate names. The powers-of-two unit is properly a gibibyte, written GiB, while gigabyte or GB refers to the decimal thousand. Most versions of Windows still print GB while actually calculating in gibibytes, which is the single biggest reason the reported figure looks low. The number on the box is honest decimal; the number on screen is binary wearing a decimal label, the quirk the overview on how drive sizes are quoted sets out plainly.
Why the two unit systems exist
Neither side is being sneaky; the split is historical. Early computing latched onto powers of two because memory addressing works in binary, so 1,024 felt natural to programmers. Storage and networking, measured and sold by the piece, kept the decimal thousand that the rest of science uses. Both conventions were entrenched long before anyone tried to reconcile them, and the gibibyte naming came late, so software rarely bothered to adopt it. The practical result is a fixed ratio: the larger the drive, the larger the apparent shortfall, though the percentage stays about the same, near seven percent at every size.
What formatting sets aside
Beyond the units, the drive has to be organised before it can hold files, and that structure costs space. Formatting lays down a file system, an index that tracks where every file begins, ends and fragments across the disk. That index, along with reserved boot areas and bookkeeping tables, occupies real capacity from the very start. The amount varies by the file system in use, a choice the article on how file systems organise a disk explains, and on larger drives it is a small slice, but it is never zero.
The file system also stores files in fixed blocks called allocation units, often four kilobytes each. A file smaller than one block still claims a whole block, so the leftover room in that block, known as slack, cannot hold anything else. Across thousands of tiny files the slack adds up, which is why a drive full of small documents shows less free space than the sum of its file sizes alone would suggest.
Recovery partitions and system files
On a prebuilt laptop or desktop the visible drive is often smaller still, because the manufacturer carves out hidden partitions. A recovery partition holds a factory image so the machine can be reset without a download, and it can take several gigabytes. The operating system then adds its own residents: a page or swap file, a hibernation file often sized to match the installed memory, system restore snapshots, and a reserved area for updates. None of these show as ordinary files, yet all of them consume capacity that never appears as free space.
Keeping the operating system on one drive and personal files on another stops these reservations from eating into your document space, a habit the article on separating system and personal storage makes the practical case for.
Solid-state drives hold back extra room
Solid-state drives introduce one more subtraction. A portion of the flash is fenced off as spare area the drive uses internally to manage wear and replace failing cells, a practice the explainer on reserved space inside an SSD describes. This over-provisioning is deliberate and helps the drive last, but it means an SSD can present slightly less usable room than a mechanical drive of the same nominal size. On many consumer models the reserved portion is around seven percent of the raw flash, and some drives set aside more to lengthen their working life. It is capacity working on your behalf rather than capacity lost.
What usable space to actually expect
Putting the pieces together, here is the rough capacity a few common drive sizes report once units and overhead are accounted for.
| Advertised size | Shown as (approx) | Typical free after setup |
|---|---|---|
| 256 GB | 238 GiB | 210 to 225 GiB |
| 512 GB | 477 GiB | 440 to 465 GiB |
| 1 TB | 931 GiB | 890 to 915 GiB |
| 2 TB | 1,863 GiB | 1,790 to 1,840 GiB |
What this means in practice
The disappearing space is really two effects stacked together. Most of it is not lost at all, just relabelled by software that counts in binary while the box counts in decimal, a difference of about seven percent you can predict for any size. The rest is genuine overhead, the file system index, recovery partitions, swap and hibernation files, and on solid-state drives a slice of reserved flash, all of which buy you reliability and recoverability.
Knowing this changes how you shop. If you need a specific amount of free room, buy a size above it rather than exactly it, since a nominal one terabyte drive gives you closer to 900 usable gigabytes after the system settles in. The figure on the label is accurate for what it measures, the figure on screen is accurate for what it measures too, and the difference between them is now something you can account for instead of resent.
Common questions, answered
Why does my 1TB drive only show 931GB?
Because your operating system counts in powers of two while the manufacturer counts in powers of ten. One decimal terabyte is a trillion bytes, and dividing that by 1,024 three times gives about 931 of the binary units the system displays as GB. No space is missing; the same bytes are simply being measured and labelled by two different conventions that disagree by roughly seven percent.
Is the missing space wasted or recoverable?
Most of it was never missing, just counted differently, so there is nothing to recover there. The genuine overhead, the file system index, swap and hibernation files, and recovery partitions, is doing a job. You can reclaim some by shrinking the hibernation file or removing a recovery partition, but that trades convenience and safety for a few gigabytes, which is rarely worth the effort.
Are drive makers being misleading about capacity?
Not technically. They quote capacity in standard decimal units, where a gigabyte is exactly a billion bytes, consistent with how the wider science world measures things. The mismatch comes from operating systems that calculate in binary yet print the decimal label GB. If anything here misleads, it is that software labelling rather than the honest number printed on the drive box.
