Earning 176 million yuan a day, the 46-year-old mathematics PhD uses AI for stock trading and becomes a tycoon with a fortune of hundreds of billions.

Another tycoon with a fortune of over 100 billion RMB has emerged in the UK.

In May 2026, *The Sunday Times* released the UK Rich List. 46-year-old mathematics PhD Alex Gerko ranks 8th with a net worth of 16.006 billion pounds (equivalent to about 141.7 billion RMB).

Within one year, his wealth surged by 7.261 billion pounds, roughly 64.3 billion RMB, representing an average daily increase of about 176 million RMB.

What pushed him into the top 10 richest people in the UK is a company with only around 300 employees — XTX Markets.

This company does not sell products or manufacture chips. It makes profits by trading in global financial markets with mathematical models and AI. The capital it handles every day is as high as 250 billion US dollars, equivalent to about 1.8 trillion RMB. In 2025, its main UK business entities recorded a total profit of about 1.7 billion pounds alone.

Only 300 people support a financial company with a daily transaction volume at the trillion-level. A wealth creation process driven by algorithms and computing power is unfolding on Wall Street and in the City of London.

Mathematics PhD founded a 100-billion-level company

Gerko's story starts in Moscow.

He was born in 1979 in Moscow, the capital of Russia. Unlike most financial tycoons, he does not have a business school background. He pursued his doctorate in mathematics at Lomonosov Moscow State University, focusing on pure mathematics fields such as commutative algebra, and later obtained a master's degree in economics at the New Economic School in Russia.

In 2004, 25-year-old Gerko joined Deutsche Bank.

At that time, Wall Street was undergoing a transformation. More and more mathematicians, physicists and programmers entered financial institutions, using computer models to replace part of the work of traditional traders.

Gerko was one of them.

He initially engaged in quantitative stock trading, and later moved to the foreign exchange market. In 2009, he left Deutsche Bank and joined the UK hedge fund GSA Capital, in charge of the foreign exchange quantitative trading business. This experience brought Gerko his first access to huge wealth.

According to materials later disclosed by the UK Supreme Court, the team led by Gerko developed automated foreign exchange trading strategies by writing software. The team was originally entitled to share 35% of the profits it generated, and the profit-sharing ratio rose to 50% after the relevant business restructuring in 2010.

Gerko was also the one who decided how much each person should get.

He is not only proficient in mathematics, but also understood the core essence of quantitative finance very early: as long as the trading system is profitable enough, a small number of people who master the algorithms can get compensation unimaginable for traditional financial practitioners.

In 2015, 36-year-old Gerko left GSA Capital and founded XTX Markets. The company name XTX, which carries the mathematician's personal interest, comes from a mathematical expression.

Gerko did not follow the organizational model of traditional financial companies.

Trading institutions common on Wall Street usually have a large number of traders, analysts and sales personnel. XTX, by contrast, leaves most trading decisions to algorithms, and concentrates human resources on mathematical research, software development and system construction.

It also did not take the typical tech startup path: raising funds first, expanding the team, continuously burning money, and then seeking a public listing.

XTX remains privately held, with Gerko holding at least 75% of the equity. The company mainly accumulates capital through its own trading business, and does not need to continuously raise funds from venture capital institutions like ordinary startups.

This means that most of the profits earned by XTX ultimately remain in the hands of the founder and a small number of partners.

Ten years later, Gerko's wealth has reached 16 billion pounds.

There is a rather interesting detail. In 2024, Gerko also established a personal investment institution called Cromulon Capital, dedicated to managing his ever-expanding wealth.

In other words, he is already so wealthy that he needs to set up an additional company to manage the money he earned from XTX.

300 people generated 1.7 billion pounds in profit within one year

Just how profitable is XTX? We can check the latest financial data.

In 2025, the three main business entities of XTX in the UK recorded a total revenue of about 3.9 billion pounds, a year-on-year increase of 44%; the profit was about 1.7 billion pounds, a year-on-year increase of 33%.

A rough calculation based on the above total figures shows that the profit margin exceeds 40%.

For comparison, the total profit of these entities in 2024 was about 1.3 billion pounds. In just one year, it earned an extra 400 million pounds.

What is even more astonishing is how this money is distributed. Relevant media disclosed in August this year that Gerko obtained 895 million pounds of income from XTX in 2025, a 31% increase over the previous year.

The other 23 senior quantitative traders and technical developers shared 833 million pounds together.

After calculation, each of these 23 people gets an average of about 36.22 million pounds, equivalent to more than 300 million RMB.

24 people shared about 1.7 billion pounds. Of course, this is the profit distribution for core partners and senior staff, and does not mean that all employees of XTX can get such income.

But the remuneration of ordinary employees is also quite considerable. According to relevant media analysis, the per capita salary of the technology entities under XTX in 2025 was about 457,000 pounds, equivalent to more than 4 million RMB.

Where does the money come from? The answer is market making. In layman's terms, market makers are like wholesalers in the financial market.

Assume that for a stock, someone is willing to sell at 100 units of currency, and another person is willing to buy at 100.02 units of currency. The market maker can provide both bid and ask quotes at the same time, and strive to earn the spread in the middle under the premise of bearing position risk and transaction costs.

Each transaction may only earn a few cents, or even less. But if you process a huge number of orders every day, the tiny spreads can add up to considerable revenue.

The real difficulty lies in risk control. If a market maker just buys at 100 units of currency and the stock price plummets to 99 units of currency, the few cents it originally expected to earn cannot make up for the loss at all.

Therefore, market making trading is not only a competition for who can place orders faster, but also requires judging what the asset may be worth in the next second, the next minute, and even several hours later.

This is exactly Gerko's advantage. XTX uses machine learning models to continuously analyze more than 53,000 financial instruments including stocks, bonds, foreign exchange, commodities and cryptocurrencies, predict price changes, and then decide how to quote and trade.

Its business is not limited to market making, and it also includes looking for trading opportunities based on the statistical relationships between different assets.

This system has covered financial markets in 35 countries around the world, with a daily transaction volume of 250 billion US dollars.

But it is worth making a special distinction: 250 billion US dollars is the transaction amount, not the money XTX earns every day.

Its main advantage is converting extremely small unit transaction returns into profits through huge transaction volumes.

The greater the market volatility, the higher the requirements for quotation and risk control. Market makers that can accurately identify price changes often get more opportunities.

2025 is a case in point. The drastic fluctuations in the financial market triggered by US tariff policies drove the growth of transaction volume of XTX's European stock business.

According to the company's announcement, the total annual transaction value of XTX's relevant client trading platforms in the UK and Europe reached 177 billion euros in 2025, a year-on-year increase of 50%. In April alone, its average daily transaction volume exceeded 1 billion euros for the first time.

For Gerko, there is another more important wealth effect. Since XTX is not listed, his personal net worth mainly depends on the estimated value of the shares he holds, rather than the public quotation formed by the stock market every day.

Therefore, the so-called daily increase of about 176 million RMB is the average distribution of the annual wealth valuation increase, which does not mean that such a large amount of cash enters the bank account every day.

Nevertheless, the 895 million pounds he actually received in 2025 is already astonishing. Gerko has even become a major taxpayer in the UK.

The tax list released by *The Sunday Times* in early 2026 shows that his estimated tax payment for the previous year reached 331.4 million pounds, ranking second in the UK.

A team of several hundred people generated billions of pounds in profits. The profit distribution model of this business is completely different from that of traditional enterprises.

Wall Street has started to compete for computing power

Gerko's wealth story is not an isolated case.

Nowadays, among the most profitable financial institutions in the world, a group of trading companies operating relying on mathematics, algorithms and computer systems are on the rise.

Take Jane Street in the United States as an example. According to data disclosed by relevant media in April this year, Jane Street's net trading revenue in 2025 reached 39.6 billion US dollars, equivalent to about 2800 billion RMB, hitting an all-time high.

What does this mean? This quantitative trading company that does not serve ordinary consumers generates revenue from its trading business alone that has exceeded that of some large global investment banks.

In the same year, Citadel Securities under Ken Griffin also recorded a trading revenue of 12.2 billion US dollars, a year-on-year increase of 25%.

Wall Street used to make profits relying on huge trading floors. Now, more and more profits are flowing to institutions that master high-performance computing, models and data.

However, this also brings a new problem. When all competitors start to use AI, who can continue to make profits?

The answer given by XTX is: continue to invest in computing power.

At the beginning of 2025, XTX announced that it would invest more than 1 billion euros to build five data centers in the Kajaani region of Finland to provide computing resources for future machine learning research.

This is not a traditional financial center, but it has become the next strategic asset that Gerko has bet on.

According to the latest public information of the company, XTX's research computing cluster already has more than 25,000 GPUs, and the available storage capacity exceeds 1 Exabyte (EB), which is equivalent to more than 1 million TB.

More importantly, it chooses to build its own data centers instead of fully relying on external cloud computing services.

The underlying business logic is not complicated. Quantitative trading institutions need to analyze massive amounts of market data every day to train and update price prediction models.

The more complex the model is, the more computing resources it consumes. If you rely on external computing power suppliers for a long time, you will not only bear the rental cost, but also may be affected by the shortage of GPU resources.

Building your own computing power infrastructure gives you the opportunity to reduce long-term training costs and keep core research capabilities in your own hands.

As of September 2026, Finnish construction company YIT has announced that it has won a further construction contract for the third facility of XTX's Kajaani data center park, covering interior decoration, electromechanical systems and commissioning. This project worth more than 1 billion euros is being implemented step by step.

It is worth noting that XTX is not exactly the same as traditional high-frequency trading institutions. In the past, a large part of the competition in the high-frequency trading industry focused on speed.

In order to be a few microseconds faster than their opponents, trading institutions did not hesitate to spend money to build dedicated communication lines and place servers close to exchanges.

But XTX puts more emphasis on the accuracy of price prediction and the ability to bear position risks. Speed is still important, but the importance of computing power, model quality and data accumulation is on the rise.

This also explains why XTX is willing to invest more than 1 billion euros to build computing power facilities. What it buys is not ordinary office equipment, but the infrastructure for future trading profits.

There are similar stories in China. Liang Wenfeng, the founder of DeepSeek, got his earliest wealth from High-Flyer Quant, a quantitative investment institution.

In 2015, Liang Wenfeng co-founded High-Flyer, and then invested heavily in machine learning and computing resources.

In 2020, High-Flyer built a computing cluster with 1100 NVIDIA A100 chips. In 2021, it built another cluster with a scale of about 10,000 A100 chips.

These investments not only served financial transactions, but also accumulated technical and talent foundations for DeepSeek's artificial intelligence research later.

The quantitative business itself also generated considerable returns. Bloomberg cited third-party statistics in January this year that the average return rate of the funds under High-Flyer in 2025 reached 56.6%, ranking among the top of the 10-billion-level quantitative institutions in China.

Of course, High-Flyer and XTX are not exactly the same companies. High-Flyer mainly operates quantitative investment funds and earns investment returns and related fees; XTX focuses on proprietary trading and providing market liquidity.

But they have one thing in common: the cash flow generated by financial transactions is supporting expensive AI R&D; and the advancement of AI technology may in turn improve the competitiveness of trading institutions.

This cycle is also accompanied by new risks. When more and more institutions use similar data and models, trading strategies that used to be profitable may quickly become invalid. Once extreme market fluctuations occur, the mutual influence between algorithms may also amplify risks.

This business is not guaranteed to make a profit. But at least for now, leading companies with strong computing resources, high-quality data and research talents are taking the upper hand.

For Gerko, the investment direction after making money has also gone beyond the financial field.

According to reports from relevant media, he and XTX have also invested in AI company Anthropic, as well as UK autonomous driving company Wayve.

XTX's official website shows that since 2020, the company has pledged to donate more than 400 million pounds to education, scientific research and other public welfare projects, with special emphasis on the cultivation of mathematical talents.

Gerko and his wife also supported the establishment of a school focusing on mathematics education. From mathematical research, to quantitative trading, to building AI data centers and investing in cutting-edge technology companies, Gerko's wealth has been continuously concentrated in the same direction: mathematics, algorithms and computing power.

More than 20 years ago, he was just a mathematics PhD who joined Deutsche Bank. Now, with a net worth of 16 billion pounds, he has ranked among the top 10 richest people in the UK. And his company has only about 300 employees.

This is probably a highly representative business landscape in the AI era: creating huge wealth no longer necessarily requires a huge workforce, but may require tens of thousands of GPUs, and a very small number of people who can harness these computing power.

This article is from the WeChat official account "Pencil News" (ID: pencilnews), written by Pencil News, and published with authorization from 36Kr.