How a Moscow-born mathematician turned AI into a money-printing machine - WSJ

Alex Gerko's XTX Markets will start using its first in-house data center in Finland this year / Photo: Shutterstock.com
Alex Gerko, founder of British trading company XTX Markets, has become one of the richest people in the UK, according to The Wall Street Journal. According to Forbes, Gerko's fortune grew by 66% over the year to $12 billion, which ensured him seventh place in the list of British billionaires.
Herco owns at least 75% of XTX Markets, WSJ notes, citing company documents. According to them, in 2025, the revenue of the British division of XTX increased by 44% to $5.3 billion, profit increased by 33% to $2.3 billion. The figures do not take into account the results of the Singapore office, which has not yet reported for last year, but previously received comparable revenue with the British division, writes WSJ.
Gerco is preparing to double its bets this year, according to the article. XTX Markets is expected to start using its own data center in Finland this year. Its goal is to expand the company's ability to process financial data and train models to predict price movements.
How does AI help Herco make money?
Long before everyone knew about ChatGPT, Herco created an artificial intelligence system aimed solely at one goal: making money, writes WSJ. XTX has about 250 employees - that's a pretty compact staff. They manage an army of algorithms based on deep learning models - that's the name given to neural networks modeled after the human brain and designed to solve complex problems. With their help, traders can predict changes in asset prices with the accuracy of milliseconds, minutes or hours.
Some XTX trades are related to market making, a strategy in which an investor is willing to both buy and sell an asset and capitalize on the spread between the bid and ask price. Other trades are based on identifying deviations from standard market patterns and betting that prices will return to normal levels.
XTX makes an average of $250 billion in trades per day, including in stocks, bonds, derivatives, currencies and cryptocurrencies. According to WSJ sources, the company recently began trading electricity.
XTX uses 25,000 artificial intelligence chips, mostly developed by Nvidia, to train its models. The company is investing more than $1 billion to build a complex of five data processing centers in Finland - the location of the site will help prevent systems from overheating. The second of the five facilities is currently being completed there and is scheduled to be operational in 2027.
XTX is building the complex in Finland on its own, rather than leasing computing capacity as is common in the financial industry, the WSJ notes. The intention is to provide the company with the resources to train models inexpensively, even as artificial intelligence booms and takes up more and more data center capacity around the world.
"The approach they appear to have taken is sort of a big bet va-bank on deep learning," Agustin Lebron, a senior researcher at Prague-based trading firm EquiLibre Technologies, told the publication.
What is Herko known for
Herko is 46 years old. He was born in Moscow and received his PhD in mathematics from Moscow State University. He moved to the UK in 2006. Gerko started out as a trader at Deutsche Bank, then went to hedge fund GSA Capital, where he managed currency trading, before founding XTX in 2015. Gerko holds UK citizenship, he has renounced his Russian citizenship.
As a philanthropist, Gerko supports astronomy research and math education, the WSJ writes. One of his favorite projects is the 1729 School for Mathematically Gifted Children, set to open in London in September. "I've maintained an interest in math education because it was a path from a, shall we say, dysfunctional past to a very successful career," Herko himself said in 2023.
This article was AI-translated and verified by a human editor
