| Gibioctets (Gio) | Yottabits (Ybit) |
|---|---|
| 1 Gibioctet | 8.589934592 × 10-15 Ybit |
| 2 Gibioctets | 1.7179869184 × 10-14 Ybit |
| 3 Gibioctets | 2.5769803776 × 10-14 Ybit |
| 4 Gibioctets | 3.4359738368 × 10-14 Ybit |
| 5 Gibioctets | 4.294967296 × 10-14 Ybit |
| 10 Gibioctets | 8.589934592 × 10-14 Ybit |
| 20 Gibioctets | 1.7179869184 × 10-13 Ybit |
| 25 Gibioctets | 2.147483648 × 10-13 Ybit |
| 50 Gibioctets | 4.294967296 × 10-13 Ybit |
| 100 Gibioctets | 8.589934592 × 10-13 Ybit |
| Reference | Gibioctets (Gio) | Yottabits (Ybit) |
|---|---|---|
| A plain text message (160 characters) | 0.000000149012 Gio | 1.28 × 10-21 Ybit |
| A three-minute MP3 | 0.00279397 Gio | 2.4 × 10-17 Ybit |
| A smartphone photo | 0.00372529 Gio | 3.2 × 10-17 Ybit |
| A high-definition film | 3.72529 Gio | 3.2 × 10-14 Ybit |
| A dual-layer Blu-ray disc | 46.5661 Gio | 4 × 10-13 Ybit |
The gibioctet is a unit of digital information equal to 1,073,741,824 octets, which is 1,024 mebioctets. Its symbol is Gio. It is the binary counterpart of the gigaoctet, and the 7.4 per cent difference between them is the single most common cause of confusion about data units.
This is the unit in which computer memory is actually sold, whatever the packaging says. A module labelled 8 GB holds 8 gibioctets, which is 8.59 gigaoctets. It could not hold anything else: memory addressing is binary, and a module with a decimal capacity would leave part of its address range unpopulated. The label is a convenient shorthand rather than a measurement.
Storage, by contrast, is genuinely decimal. A drive sold as 500 gigaoctets holds five hundred thousand million octets, and the operating system divides that by 1,024 three times and reports 465.7 gibioctets. The two figures describe the same hardware. Neither party is being dishonest; they are simply counting in different bases, and the labelling convention differs by industry.
Cloud computing has made the distinction visible in contracts. Virtual machine sizes, container memory limits and the parameters that control how much memory a program may use are all specified in gibioctets, because they map onto real memory pages. A limit of 4 GB set in a decimal-minded tool and 4 GiB set in a binary one differ by nearly three hundred megaoctets, which is enough to change whether a program runs.
For everyday scale, a gibioctet holds about a thousand books as plain text, a few hundred photographs, or twenty minutes of high-definition video. It is roughly what a modern web browser occupies in memory with a dozen pages open, which is a fair illustration of how the standard for what counts as a small amount has moved.
The correct symbol is Gio, and it appears in system tools, in virtualisation documentation and in standards. Where a document writes GB for memory, the safe reading is gibioctets; where it writes GB for a disc or a network, the safe reading is gigaoctets. The habit of checking which is meant costs a second and prevents a seven per cent error.
One gibioctet equals 1,073,741,824 octets, 1,024 mebioctets, 8 gibibits, or about 1.074 gigaoctets.
The yottabit is a unit of digital information equal to a thousand zettabits — a one followed by twenty-four zeros. Its symbol is Ybit. For thirty years it was the largest unit the metric system offered, and it describes a quantity of data that does not yet exist anywhere.
A yottabit is 125 zettaoctets. The total of all data held by humanity is currently estimated in the low hundreds of zettaoctets, so the world's entire stock of information is somewhere around one or two yottabits. It is the first data unit that is genuinely larger than the thing it might describe, which is what makes it interesting.
The prefix yotta was adopted in 1991 at the 19th General Conference on Weights and Measures, alongside zetta. The names were invented rather than derived: the Greek and Latin numerals had been used up by peta and exa, and the committee needed letters that were not already taken as unit symbols. Y and Z were among the few remaining, and the syllables were built around them.
For a long time the yotta prefix marked the top of the ladder, and it appeared mainly in speculation and in claims about intelligence agencies' storage plans that turned out to be exaggerated. Then in 2022 the General Conference added ronna and quetta above it, together with ronto and quecto below, because data quantities were rising fast enough that the top was in sight.
Reaching a yottaoctet of storage with current technology would be a physical undertaking. Even at the highest densities available, it would require hundreds of millions of the largest hard drives ever made, a power supply comparable to that of a small country, and a building programme lasting decades. The constraint is not the mathematics but the silicon and the electricity.
The unit is nevertheless properly defined and a converter must handle it, both because forecasts of future data growth are written in yottaoctets and because the metric system's rule is that every prefix applies to every unit. A quantity does not need to exist for its name to be well formed.
One yottabit equals 1,000 zettabits, 125 zettaoctets, or about 0.8272 yobibits.