Free Data Storage Unit Converter Online

Bits vs bytes — the foundation of digital storage

Every piece of digital information ultimately comes down to binary digits. A bit (binary digit) is the smallest possible unit of digital information — it holds either 0 or 1. Eight bits make one byte, which can represent 256 distinct values (2⁸). Almost all storage capacity, file sizes, and memory specifications are expressed in bytes or their multiples.

The bit/byte distinction matters enormously in practice. Network speeds are specified in bits per second (Mbps, Gbps) because network hardware transmits data one bit at a time. File sizes and storage are in bytes. A "100 Mbps" internet connection delivers 100 million bits per second — divide by 8 to get about 12.5 MB/s maximum file download speed. This discrepancy is a frequent source of frustration: an internet plan advertised as "1 Gbps" downloads a 1 GB file in about 8 seconds in ideal conditions, not 1 second.

Network hardware and internet service providers universally use bits; operating systems, cloud storage dashboards, and file managers universally use bytes. The industry convention uses a lowercase "b" for bits and an uppercase "B" for bytes. A Wi-Fi router spec listing "Wi-Fi 6 — 9.6 Gbps" means 9.6 gigabits per second — roughly 1.2 GB/s of raw throughput before protocol overhead. Always check the capitalisation when comparing specs.

SI (decimal) vs IEC (binary) units — the MB vs MiB confusion

There are two competing and simultaneously-in-use systems for measuring data storage, and understanding the difference is essential for anyone who works with computers. The conflict arises from two different interpretations of the same prefixes — a problem that has persisted for decades and still causes confusion on a daily basis.

SI (decimal) units — base 1000

The International System of Units (SI) defines kilo as 10³ = 1,000, mega as 10⁶ = 1,000,000, and so on in powers of 10. Hard drive manufacturers, SSD makers, and USB flash drive manufacturers use SI definitions because they make their products' capacities appear larger. A "1 TB" drive contains exactly 1,000,000,000,000 bytes by this definition. Cloud providers (AWS, Google Cloud, Azure) bill storage in SI units. Networking bandwidth is also expressed in SI units: a 1 Gbps port transmits 1,000,000,000 bits per second.

IEC (binary) units — base 1024

In 1998, the International Electrotechnical Commission (IEC) introduced unambiguous binary prefixes to resolve the confusion: kibi (KiB) = 2¹⁰ = 1,024 bytes, mebi (MiB) = 2²⁰ = 1,048,576 bytes, gibi (GiB) = 2³⁰ = 1,073,741,824 bytes, and so on. These are precisely defined for binary computing. RAM, CPU caches, virtual memory pages, and file system allocation blocks are all naturally measured in powers of 2, so IEC units are the correct choice for these. Windows has always displayed storage in binary units but labelled them incorrectly as "MB/GB." macOS switched to SI units in 10.6 Snow Leopard (2009), eliminating its own "missing space" issue.

Notice how the difference grows with scale: at the kilobyte level, the discrepancy is only 2.4%, but by the time you reach petabytes, IEC units are 12.6% larger than their SI counterparts. At the exabyte level (a common unit for hyperscale data centres), the difference exceeds 15%.

Why does a "1 TB" drive show only 931 GB in Windows?

This is one of the most frequently asked questions in consumer technology, and the answer requires understanding the SI vs binary distinction above. A hard drive labelled "1 TB" contains exactly 1,000,000,000,000 bytes using the SI decimal definition that the manufacturer (correctly) applied. When Windows Explorer displays the drive capacity, it performs the division using binary units:

1,000,000,000,000 bytes ÷ 1,073,741,824 (1 GiB) ≈ 931.32

Windows then labels this as "931 GB" — using the SI symbol for a binary quantity. So Windows is showing you the correct number of gibibytes (GiB), but mislabelling them as gigabytes (GB). The drive is not smaller than advertised; there is no "missing" space — it is a unit labelling inconsistency that has persisted since the 1990s. macOS avoids this confusion entirely by displaying storage in SI units: the same 1 TB drive shows as "1.0 TB" in macOS Finder.

The practical takeaway: when comparing storage capacities, always check what OS is being used to measure. A 512 GB SSD shows as 477 GiB (≈477 "GB" in Windows) and 512 GB in macOS. For cloud storage, AWS and Azure quote in SI units (GB = 10⁹ bytes) and bill accordingly.

RAM vs storage — why they're measured differently

RAM (Random Access Memory) and persistent storage (SSDs, HDDs) serve very different roles, and they have traditionally been measured using different conventions — RAM in binary, storage in whatever the manufacturer chooses.

RAM is always binary. A "16 GB RAM" module contains exactly 2³⁴ = 17,179,869,184 bytes (16 GiB). RAM chips are manufactured in power-of-2 capacities by necessity — memory controllers and addressing hardware are built around binary arithmetic. You will never see a 16.384 GB RAM stick; the 4-byte difference between 16 GB (SI) and 16 GiB (IEC) would require non-standard addressing logic.

Persistent storage is marketed in SI. Hard drives, SSDs, and flash storage are fabricated in layouts that don't inherently require power-of-2 capacities, so manufacturers adopted SI units to present larger-looking numbers. A 1 TB NVMe SSD contains 1,000,000,000,000 bytes — exactly 1 TB in SI, but only 931 GiB in binary.

CPU caches are binary by construction. L1, L2, and L3 CPU caches are always specified in KiB or MiB: a CPU with "12 MB L3 cache" typically has 12 × 1,048,576 = 12,582,912 bytes of cache — though some manufacturers have started using SI here too, adding ambiguity.

Network speeds and file transfer — converting Mbps to MB/s

Understanding the relationship between advertised network speed (in bits) and actual file transfer rate (in bytes) is essential for developers building data pipelines, selecting cloud storage tiers, or debugging slow uploads.

Realistic throughput is lower than the theoretical maximum due to TCP/IP overhead (typically 3–5%), protocol headers, retransmission, and network congestion. For production systems, a safe rule of thumb is to budget for 70–80% of the advertised link speed for sustained bulk transfers.

Data storage at scale — from kilobytes to zettabytes

Understanding the scale of modern data requires anchoring abstract units to real-world examples. The gap between a kilobyte and a zettabyte spans 18 orders of magnitude — a range comparable to the difference between the width of a human hair and the distance from Earth to the Sun.

To put hyperscale in context: in 2024, global internet traffic exceeds 400 exabytes per month. The entire text of all Wikipedia articles in all languages is roughly 22 GB — a trivially small fraction of a single data centre. Google's total storage capacity across all data centres is estimated in the tens of exabytes. The human genome — 3.2 billion base pairs — encodes to about 800 MB in raw form and about 1.5 GB in a practical VCF file format.