A terabyte (TB) is a unit of digital data equal to one trillion bytes, or 1,000 gigabytes (GB), as defined by the international system of units (SI). Its binary counterpart, the tebibyte (TiB), is 2^40 bytes, about 10% more. The terabyte sits below the petabyte (PB) and is the everyday unit for drives, storage arrays, file servers and backup capacity, which makes it the unit where raw, usable and effective capacity are most often confused.
At a glance
- Decimal: 1 TB = 10^12 bytes = 1,000 GB. Binary counterpart: 1 tebibyte (TiB) = 2^40 bytes = 1,099,511,627,776 bytes.
- The gap between the unit and its binary counterpart is about 10%; a 1 TB drive shows as about 931 GB in Windows.
- As a rough illustration, a laptop drive, a department file share or a small company’s total file data is often in the low terabytes.
- Storage quotes may state raw, usable or effective TB, and the three can differ widely.
- Backup capacity is usually sized in TB, including retained versions, not just today’s data.
What problem it solves
Once data grows beyond a few hundred gigabytes, the terabyte becomes the natural unit for planning. Storage arrays, network-attached storage (NAS) appliances, storage area networks (SANs) and backup repositories are specified and priced in TB, and capacity planning, refresh budgets and cloud migration estimates are usually done at this scale.
The trouble is that “100 TB” can mean very different amounts of space your applications can actually use, depending on whether it really means TB or TiB and whether it is raw, usable or effective.
How it works
Decimal and binary values. Drive makers sell decimal TB. Some operating systems show binary (TiB, GiB) values under SI labels (TB, GB), so the same drive appears smaller: about 931 GB, or 0.91 TB.
Raw capacity is the sum of the drives’ labelled capacities.
Usable capacity is what remains after data protection (mirroring, RAID parity or erasure coding), hot spares, file system and metadata overhead, and space some systems reserve for snapshots or performance. The ratio of usable to raw depends on the design and the vendor.
Effective capacity applies an assumed data reduction ratio from deduplication and compression. Vendors often quote it prominently. Real ratios vary widely by data type: virtual machine images and repeated backups often reduce well; encrypted data, video and already-compressed files usually do not.
Backup sizing. Backup storage depends on the size of protected data, the daily change rate, how long versions are kept, and the reduction the backup software achieves. A backup as a service (BaaS) quote may be priced on protected TB (the source data) or stored TB (what the backups consume), which give different totals.
Moving terabytes. As a rough calculation, 1 TB takes at least about 2.2 hours at a full 1 Gbps. Migrating tens of TB into or out of a cloud can take days over the network, and leaving a cloud may incur data egress fees; providers often offer physical transfer devices for large moves.
See our file and object storage, backup as a service and colocation pages for how we help buyers size and compare these options.
When it matters for buyers
- Storage refreshes. Compare array, NAS and storage as a service (STaaS) quotes on usable TB at the same protection level.
- Backup contracts. Check whether the price is per protected or stored TB and how retention affects growth.
- Cloud migrations. Total TB and network speed together set the migration timeline.
- Growth commitments. Capacity-based contracts often have committed minimums in TB and overage rates above them.
Questions to ask vendors
- Is this quote in raw, usable or effective TB, and decimal or binary units?
- What protection scheme and spare capacity are assumed in the usable figure?
- What data reduction ratio is assumed, and is it guaranteed in writing, with what remedy if it is not achieved?
- For backup, is pricing based on protected data or stored data, and how do retention changes affect it?
- What does it cost, and how long does it take, to get our terabytes back out at the end of the contract?
How it differs from a tebibyte (TiB)
A tebibyte is 2^40 bytes, about 10% more than a decimal terabyte. On a 100 TB purchase that is a gap of roughly 10 TB, so it is worth knowing which unit a quote, a monitoring tool and a cloud bill each use. The glossary term Binary Prefixes covers the binary units in more detail.