High-frequency trading uses computer algorithms to automate trading and replace the role that humans once had in the market. They earn a small profit from the spread on a trade. In most cases, that amounts to only a few cents at best.
There are many exchanges, and they compete on offering the best price. When one price updates, they all update, so those prices come and go very, very quickly. Modern trading can take place in barely comprehendible measurements of time. Banks and high-frequency traders often boast of executing orders on stocks and futures exchanges in milliseconds or thousandths of a second.
Let's put this into perspective. The average time it takes the human finger to click the mouse is allegedly 150,000 microseconds. To be the fastest, they must use super fast computers and telecoms cables and even put their IT servers in the same room in the data centre as the exchange. For the exchange, this rent is a nice little sideline.
Not all high-frequency trading companies are the same. Some act as market makers, buying and selling other people's trades. Others use their own capital to earn a higher profit. But in general, they tend to be small companies, and the big ones often they have no more than a couple of hundred employees at most.
Some depend purely on speed and love to iron out any possible inefficiency in getting a signal from their IT equipment to the exchange. That could even include finding the fastest geographical route. Some have built huge masts several hundred feet high to ping their signals via radio from one city to another.
Others rely on crunching more data, or using different data sources to steal a march on rivals. In an attempt to stay on top of such speedy activity, the European Union is trying to synchronise the clocks on the computers that timestamp trades. What they hope to do is if something happens like a sudden market crash, they can get a moment by moment snapshot of the critical time, much like the slow motion replay of a controversial sports incident.
But it is not as profitable a business as it once was. The boom years were in 2008 and 2009, when the market volatility on slower trading systems of banks and the high-frequency traders were sometimes seconds ahead of rivals. Now, rivals have caught up, and even some of the biggest are either leading the market, merging, or being bought. For them, life may have been fast, but it was also short.