| Yottabits per second (Ybit/s) | Tebioctets per second (Tio/s) |
|---|---|
| 1 Yottabit per second | 113686837722 Tio/s |
| 2 Yottabits per second | 227373675443 Tio/s |
| 3 Yottabits per second | 341060513165 Tio/s |
| 4 Yottabits per second | 454747350886 Tio/s |
| 5 Yottabits per second | 568434188608 Tio/s |
| 10 Yottabits per second | 1136868377216 Tio/s |
| 20 Yottabits per second | 2273736754432 Tio/s |
| 25 Yottabits per second | 2842170943040 Tio/s |
| 50 Yottabits per second | 5684341886081 Tio/s |
| 100 Yottabits per second | 11368683772162 Tio/s |
| Reference | Yottabits per second (Ybit/s) | Tebioctets per second (Tio/s) |
|---|---|---|
| A dial-up modem | 5.6 × 10-20 Ybit/s | 0.00000000636646 Tio/s |
| Typical home broadband | 1 × 10-16 Ybit/s | 0.0000113687 Tio/s |
| Gigabit Ethernet | 1 × 10-15 Ybit/s | 0.000113687 Tio/s |
| Streaming a 4K film | 2.5 × 10-17 Ybit/s | 0.00000284217 Tio/s |
The yottabit per second is a unit of data transfer rate equal to a thousand zettabits per second. Its symbol is Ybit/s. For thirty years it was the largest rate the metric system could name, and it describes a speed that has no application anywhere: not in engineering, not in research, and not in any forecast that anyone takes seriously.
Its distance from reality is easy to state. Global internet traffic averages roughly one exabit per second, so a yottabit per second is about a million times the combined communication of the entire human species. The largest single link ever built runs at a few hundred terabits per second, which is a ten-billionth of this figure.
The unit is nonetheless properly defined, and that completeness is the point. The metric system's rule is that every prefix combines with every unit without exception, so a reader who has never seen Ybit/s can decode it from the prefix alone. A system with gaps would need a table of which combinations are legal, and a table is exactly what a rule-based system exists to avoid.
There is a physical way to think about the number. A yottabit per second is 125 zettaoctets per second, and the total quantity of data humanity has ever stored is a few hundred zettaoctets. A link at this rate would therefore transmit the entire accumulated information of the species in a couple of seconds. No such body of data exists in one place to be sent, and nothing at the far end could receive it.
The prefix yotta was adopted in 1991 alongside zetta, at a conference that also had to consider which letters remained free. It stood at the top of the metric ladder until 2022, when ronna and quetta were added above it, partly because data quantities were beginning to approach the old ceiling. Rates have not followed quantities upward at the same pace.
For a converter, handling the unit is a matter of consistency rather than utility. A tool that converts every metric prefix correctly does not need to decide which of them anyone will use, and a figure written in yottabits per second — however unlikely — converts by the same rule as any other.
One yottabit per second equals 1,000 zettabits per second, 125 zettaoctets per second, or about 0.8272 yobibits per second.
The tebioctet per second is a unit of data transfer rate equal to 1,024 gibioctets per second, or two to the fortieth power octets per second. Its symbol is Tio/s. It is the binary counterpart of the teraoctet per second, and the two differ by 10 per cent.
Nothing outside a large machine moves data this quickly. The rate describes the memory bandwidth of an accelerator with stacked memory, the internal fabric of a high-end processor package, or the aggregate throughput of a parallel filesystem spread across thousands of drives. All of these are built from binary structures, and their totals are binary quantities divided by time.
High-performance computing is where the unit is written most often. A supercomputer's storage system is specified by how many tebioctets per second it can deliver to the compute nodes, because that number determines how quickly a simulation can save its state and resume. A machine that computes quickly but writes slowly spends its time waiting.
The ten per cent difference from the decimal unit is significant in that context. A filesystem procured to deliver 10 teraoctets per second and one delivering 10 tebioctets per second differ by a whole teraoctet per second, which in a facility of that size represents a substantial fraction of the hardware budget.
In bits a tebioctet per second is 8 tebibits per second, and in decimal terms about 1.1 teraoctets per second. Expressing the same rate four different ways is routine at this level, because the storage industry, the memory industry, the network industry and the standards bodies each prefer a different one.
For everyday comparison, a tebioctet per second would fill a large consumer hard drive in about twenty seconds. No external interface carries this; the figure describes movement between components inside a single system, where the wires are short and there are very many of them running in parallel.
Graphics processors have brought the rate within reach of a single component. A stack of high-bandwidth memory bonded directly to the processor die delivers well over a tebioctet per second to the chip that uses it, and a card carrying several such stacks passes a few. That bandwidth, rather than raw arithmetic speed, is what limits the training of large models: the arithmetic units sit idle unless the memory can keep them fed. The same reasoning explains why supercomputer designers spend as much effort on the paths between memory and processor as on the processors themselves, and why the rate is quoted in binary units when the memory it describes is addressed in powers of two.
One tebioctet per second equals 1,024 gibioctets per second, 1,099,511,627,776 octets per second, or about 1.100 teraoctets per second.