Quadrillion

A Quadrillion is a specific, very large number. In German, it denotes a one followed by 24 zeros. It is written for short as 10^24, spoken “ten to the power of 24”. That means: the number ten is multiplied by itself 24 times. For comparison: a million has six zeros, a billion (German: Milliarde) has nine. A Quadrillion is therefore not a little bit bigger, but on a completely different order of magnitude. It is a trillion trillions.

The translation error that’s off by a factor of a billion

German and English name large numbers according to different systems. The German system is called the “long scale”, the English one the “short scale”. In German, a Quadrillion stands for 10^24. In English, “quadrillion” stands for 10^15. Between the two lies a factor of one billion. Anyone who translates an English report word for word doesn’t miscalculate by a few percent, but massively.

The same trap already appears one level earlier. The English “billion” is the German “Milliarde”, not the German “Billion”. That’s why figures in translated tech and financial reports are a well-known source of error. In serious texts, the power of ten is therefore often additionally given in brackets. That is unambiguous and works in every language.

For readers of tech news, this is more than a matter of language. The computing power, data volumes, and training effort of AI systems are today given in precisely these orders of magnitude. Anyone who misunderstands the unit will mistake an ordinary supercomputer for a sensation. Or the other way around.

How to make sense of 24 zeros in the first place

The German system follows a fixed rule. Each level ending in “-illion” is a million times larger than the previous one. Million is 10^6, Billion (trillion) is 10^12, Trillion (quintillion) is 10^18, Quadrillion (septillion) is 10^24. In between lie the intermediate levels ending in “-illiarde”: Milliarde (billion), Billiarde (quadrillion), Trilliarde (sextillion). Each of these is a thousand times larger than the level below it.

In practice, nobody calculates with zeros written out. Instead, one uses powers of ten. The advantage: multiplying becomes adding. A trillion (10^12) times a trillion (10^12) is 10^12 times 10^12, i.e. 10^24 — a Quadrillion. You simply add the exponents.

An analogy helps to get a feel for the size. A grain of sand has a diameter of roughly half a millimeter. Estimates for all grains of sand on all the beaches of the Earth lie in the range of 10^21 to 10^24. A Quadrillion is therefore roughly on the order of magnitude of “all the sand in the world”. Such numbers elude our intuition, which is why calculation rules are needed instead of imagination.

Where such orders of magnitude show up in technology

In the AI industry, the effort required to train a model is measured in FLOPs. FLOP stands for a single floating-point calculation operation. Large language models today require on the order of 10^25 to 10^26 FLOPs for their training. The EU AI Act draws a regulatory threshold at 10^25 FLOPs. That lies exactly in this numerical range, just a few levels above a Quadrillion.

This order of magnitude also appears with data volumes. A yottabyte is 10^24 bytes, i.e. a Quadrillion bytes. So far, this unit is rather theoretical. The total amount of data generated worldwide each year is measured in zettabytes, i.e. 10^21 bytes. Data centers and storage manufacturers are thus slowly moving toward the next level.

A common misunderstanding in headlines: “a quadrillion computing operations per second” sounds like a number nobody can put into context. Usually, what’s meant is the English meaning of 10^15, i.e. a petaflop. Computers of this kind have existed since 2008. Anyone who checks the unit immediately recognizes whether a report really describes a breakthrough or was simply poorly translated.