Why your hard drive shows less space than advertised

Buy a 1 TB drive, plug it in, and Windows will tell you it holds about 931 GB. That missing 69 GB is the single most common "is my computer broken?" question in consumer computing, and it has a genuinely interesting answer: the drive holds exactly what the box says. The number on the box and the number on your screen are measured with different-sized units.

Two definitions of a gigabyte

Computers count in binary, so the natural groupings for memory are powers of two: 2, 4, 8, 16, and so on up to 1,024. Early computing borrowed the metric prefixes — kilo, mega, giga — and applied them to the nearest convenient power of two. A kilobyte became 1,024 bytes rather than 1,000, because 1,024 is 2¹⁰ and lands close enough to a thousand to feel reasonable.

Storage manufacturers, meanwhile, use the prefixes the way the rest of the metric world does. To a drive maker, a kilobyte is 1,000 bytes, a megabyte is 1,000,000, and a terabyte is exactly 1,000,000,000,000. This is not marketing sleight of hand — it is the standard SI meaning of "tera," and it is arguably the more defensible reading.

The two definitions agree at small scales and drift apart as the numbers grow, because the error compounds at every step of the ladder:

Now the arithmetic works out. Your drive genuinely contains 1,000,000,000,000 bytes. Windows divides that by 1,073,741,824 to express it in binary gigabytes, and gets 931.32. Same drive, same bytes, different divisor. Nothing has been lost.

Notice too that the discrepancy grows with capacity. On an old 500 MB drive the gap was under 5% and nobody noticed. At 1 TB it is 10%, and at 16 TB you appear to be missing well over a terabyte — which is why the complaint has become much louder as drives have grown.

The units that were supposed to fix this

In 1998 the International Electrotechnical Commission proposed a clean solution: keep the metric prefixes for powers of ten, and introduce new ones for powers of two. Kibibyte (KiB) for 1,024 bytes, mebibyte (MiB) for 1,048,576, gibibyte (GiB), tebibyte (TiB), and so on.

Technically this settles the ambiguity completely. In practice adoption has been patchy. Linux distributions and some developer tooling use the IEC units properly. Windows still labels binary quantities "GB". macOS took a different route in 2009, switching to decimal units so that a 1 TB drive displays as roughly 1 TB — which is why the same external drive can show noticeably different capacities on a Mac and a PC sitting side by side.

So if you have ever compared file sizes across two machines and found they disagreed, this is usually why. Neither is wrong; they are using different units under the same label.

Where the rest of the space goes

Binary-versus-decimal explains the large, predictable gap. It does not explain everything — if you look at a freshly formatted drive you will typically find a bit less free space even than 931 GB. Three things account for the remainder.

The file system itself takes room. A drive needs structure before it can hold files: allocation tables, journals, index metadata. NTFS reserves space for the Master File Table; APFS and ext4 maintain their own metadata. On a large modern drive this is well under 1%, but it is not nothing.

Slack space rounds every file up. Storage is allocated in fixed clusters — commonly 4 KB. A 1 KB text file still occupies a full 4 KB cluster, wasting 3 KB. One file is irrelevant; a project folder with 50,000 small files can lose several hundred megabytes to rounding. This is why "size" and "size on disk" differ in file properties.

SSDs hold spare capacity in reserve. Solid-state drives keep a pool of blocks invisible to the operating system, used for wear levelling and to swap in when blocks fail. This over-provisioning is deliberate and it is what keeps the drive fast and reliable as it ages — some manufacturers reserve around 7% for it.

There is also a recovery partition on most pre-built machines, typically 10–30 GB, which is counted against the drive but not shown as usable space.

Megabytes and megabits are not the same thing

A closely related confusion is worth clearing up because it costs people money. Internet plans are sold in megabits per second (Mbps). File sizes are in megabytes (MB). There are eight bits in a byte, so:

• MB/s = Mbps ÷ 8

A 100 Mbps connection downloads at a theoretical maximum of about 12.5 MB/s — so a 5 GB game takes roughly seven minutes at full speed, not under a minute. Providers quote the bigger-sounding unit; your download manager shows the smaller one. Both are accurate. The lowercase "b" versus uppercase "B" is doing a lot of work in that distinction, and it is easy to miss.

What to do about it

Mostly, nothing — the capacity is real and the drive is fine. But a few habits help:

Buy with the 7–10% gap in mind. If you need 1 TB of actual usable space on Windows, a 1 TB drive will not quite get you there once formatting and the recovery partition are accounted for. Size up.

Compare like with like. When a limit and a file size come from different places — an email attachment cap versus a file size shown by your OS, say — check whether both are using the same standard before concluding the file will fit.

Do not fill an SSD. Performance degrades noticeably as a solid-state drive approaches capacity, because the controller has fewer free blocks to work with. Keeping 10–15% free is a reasonable rule, which effectively widens the gap between the number on the box and the space you should actually plan to use.

None of this is a defect. It is two reasonable conventions, adopted for different reasons by different industries, both of which are now too entrenched to change. Once you know which divisor is in play, the numbers stop being mysterious.

Tools mentioned in this guide