The prefixes kilo, mega, giga, tera, and peta are used to express quantities of bytes in increasing orders of magnitude. Each level represents a multiplicative factor compared to the previous one, and the confusion between these units regularly leads to misjudgments when purchasing a hard drive, choosing a cloud plan, or managing large backups.
Decimal prefixes and binary prefixes: a rarely explained ambiguity
When a hard drive manufacturer displays a capacity in terabytes, they use the decimal system: each level corresponds to a power of 10. One kilobyte is therefore 1,000 bytes, one megabyte is 1,000,000 bytes, and so on. Your computer’s operating system, on the other hand, often calculates in powers of 2. One kibibyte (Ki) is worth 1,024 bytes instead of 1,000, and one mebibyte is 1,048,576 bytes.
The International Electrotechnical Commission (IEC) introduced binary prefixes (kibi, mebi, gibi, tebi, pebi) as early as 1998 to ensure that SI prefixes retain their decimal meaning. In practice, few consumer software applications have adopted this nomenclature, which perpetuates the confusion.
The gap between decimal value and binary value increases with size. For a kilobyte, the difference is negligible. At the terabyte scale, it exceeds several tens of gigabytes. A drive sold as one terabyte (decimal) therefore shows a lower capacity once formatted and read by the system, without the manufacturer having lied. Understanding the difference between kilo, giga, tera, and peta first requires knowing which reference frame is being used.
Conversion table for storage units in bytes
A table is better than a long explanation for memorizing orders of magnitude. Here is the correspondence between each prefix, its value in decimal system (SI), and its binary equivalent (IEC).
| SI Prefix | Symbol | Decimal Value | Binary Equivalent | IEC Symbol |
|---|---|---|---|---|
| Kilo | Ko | 10³ (1,000 bytes) | 2¹⁰ (1,024 bytes) | Kio |
| Mega | Mo | 10⁶ | 2²⁰ | Mio |
| Giga | Go | 10⁹ | 2³⁰ | Gio |
| Tera | To | 10¹² | 2⁴⁰ | Tio |
| Peta | Po | 10¹⁵ | 2⁵⁰ | Pio |
Each row represents a factor of 1,000 (decimal) or 1,024 (binary) compared to the previous one. Moving from one level to another multiplies the number of bytes involved by a thousand.
Kilo, mega, giga, tera, peta: what do these volumes correspond to in daily life
Raw numbers don’t speak for themselves. What matters is knowing what type of content fits into each order of magnitude.
- The kilobyte is the scale of plain text: a document of a few pages weighs a few tens of Ko. A photo, even compressed, already exceeds this level.
- The megabyte is the common unit for everyday files: a high-resolution photo weighs several Mo, a compressed music track between three and ten Mo.
- The gigabyte corresponds to the storage of applications, games, or videos. A standard definition movie occupies between one and two Go, and mobile data plans are counted in Go.
- The terabyte has become the standard for consumer hard drives and home NAS devices. It allows for archiving entire libraries of photos, videos, or system backups.
- The petabyte remains reserved for professional infrastructures: data centers, scientific archives, large-scale streaming platforms.
Moving from kilo to peta means crossing five levels of a thousand. This geometric progression explains why storage needs seem to explode: a 4K video file weighs thousands of times more than a text file, and the volumes of data produced each year are now measured on the scale of peta, or even beyond.
New SI prefixes adopted in 2022: ronna and quetta
The hierarchy does not stop at peta. After exa (10¹⁸), zetta (10²¹), and yotta (10²⁴), the General Conference on Weights and Measures adopted two new prefixes in 2022: ronna (R, 10²⁷) and quetta (Q, 10³⁰). On the binary side, the IEC nomenclature has been enriched with robi (Ri, 2⁹⁰) and quebi (Qi, 2¹⁰⁰).
This extension responds to a concrete need. The volumes of data generated by scientific research, artificial intelligence, and the Internet of Things exceed the scales that yotta could properly name. The SI resolution also aimed to prevent unofficial terms from taking hold in usage, as had happened with “hella” informally proposed for 10²⁷.
Why these prefixes matter for future storage
The largest data centers are already handling volumes approaching the zettabyte on a global scale. Having standardized prefixes beyond yotta allows engineers and standardization bodies to unambiguously designate capacities that, ten years ago, were in the realm of speculation. Each additional level multiplies capacity by a thousand, and the exponential growth of data makes these designations progressively necessary.
The distinction between decimal and binary prefixes remains valid at these scales. One ronnaoctet (10²⁷ bytes) and one robioctet (2⁹⁰ bytes) do not designate exactly the same quantity, and the gap widens even more as the exponent increases.
Retaining the logic rather than the exact numbers remains the best approach. Each prefix is a thousand times the previous one in decimal and 1,024 times in binary. Once this principle is grasped, moving from kilobyte to quettaoctet only requires counting levels.
