HomeArticle

Earning 170 million yuan per day, the 46-year-old mathematics PhD breaks into the top 10 of the wealth rankings by leveraging AI.

新智元2026-10-19 08:58
The Brit who is best at making money with AI has emerged.

Just now, the Briton who makes the most money with AI has appeared!

According to the latest 2026 UK Rich List released by The Sunday Times, 46-year-old Alex Gerko, a mathematics doctor from Lomonosov Moscow State University, has strongly broken into the top 10 of the UK's rich list and ranked 8th with an astonishing net worth of 16.006 billion pounds (about 141.7 billion yuan).

In just the past year, his wealth has soared by 64.3 billion yuan — which is equivalent to nearly 176 million yuan being automatically credited to his account every day when he opens his eyes.

His huge wealth comes from XTX Markets, an AI quantitative trading company he founded in 2015.

This company has completely broken the traditional Wall Street model. The entire company has only more than 250 employees, no traders in the traditional sense, and no external investors.

Their "money printer" is a pure AI trading system with more than 400PB of global financial market data, driven by a GPU cluster.

XTX has quickly become one of the most profitable private enterprises in the UK. As the absolute controller holding about 75% of the company's shares, Gerko takes the largest share of the wealth pie alone.

AI automated trading is subverting traditional finance.

Abandoning humans: the "money printing factory" with no traders

Gerko was born in the suburbs of Russia, graduated from the famous Faculty of Mechanics and Mathematics of Lomonosov Moscow State University for his bachelor's degree, and then obtained his doctorate in mathematics.

The mathematics training of the Russian school is known for being solid, rigorous and hardcore. In his perception, the essence of the world is not metaphysics, but probability, distribution and stochastic differential equations.

After graduating with his doctorate, Gerko successively worked at Deutsche Bank and the hedge fund GSA Capital.

It was during this period that Gerko saw through the fatal weakness of traditional trading: human beings.

Humans have emotions, they fear and they are greedy; human reaction speed has physical limits; humans cannot process thousands of variables in their brains at the same time.

So in 2015, Gerko decided to go it alone and founded XTX Markets.

The core concept of this company is extremely extreme: "fire" all human traders.

XTX has built a pure AI trading system completely for the purpose of making profits.

Among the more than 250 employees of the whole company, the vast majority are algorithm engineers, data scientists and underlying hardware geeks, and there are no "traders" in the traditional sense.

In the past, the global financial market was compared to a huge casino, ordinary investors are gamblers betting on ups and downs, then market makers are the people standing by the gambling table, ready to provide chip exchange services at any time.

What market makers earn is not the windfall after betting in the right direction, but the gap between the bid price and the ask price, which is the bid-ask spread in financial terms.

The profit from each transaction may be very small, but if you accumulate little by little, you can roll out amazing profits.

Of course, this is by no means a guaranteed profit.

The biggest risk for market makers is that the market suddenly runs wildly in one direction. For example, when the stock market plummets, everyone is in a hurry to sell, but market makers hold a large number of assets in their hands.

In the past, this business was highly dependent on top traders. They needed to constantly adjust quotations, control inventories, and hedge risks.

But Gerko completely removed human traders from trading decisions and let machines take over the entire trading process.

Mathematical models, machine learning and automated trading systems have replaced human experience and intuition. Quotation, order placement, risk management and hedging execution are all led by algorithm systems.

"In front of real mathematics and algorithms, the so-called 'trading sense', 'experience' and 'intuition' of human traders are as backward as medieval astrology."

Why can a company with 250 people crush Wall Street behemoths?

At XTX Markets, pronounced X-Transpose-X, there is a fanatical belief in pure mathematics.

In the financial world, XTX Markets is a "maverick".

First of all, there are zero traditional traders here.

Walk into XTX, all trading judgments, microsecond-level matching, and risk hedging are 100% decided independently by AI deep learning models.

This company does not recruit any traditional traders. More than 70% of its employees are mathematicians, cryptographers, machine learning experts and underlying architecture engineers.

The only mission of human employees at XTX is to write higher-dimensional mathematical formulas, debug neural network models, and maintain computing power infrastructure.

What's more ruthless is that they don't serve investors, and there is absolutely zero external capital.

Most of the hundred-billion-dollar giants on Wall Street make a living by managing other people's money and charging management fees. But XTX rejects any external capital (LP). They only use their own real money to fight in the market, so every penny they earn belongs entirely to the team, and they do not share the profits with any Wall Street giants.

In XTX Markets' London headquarters, there is a replica of the Apollo 11 command module, as well as a genuine 1943 Enigma encryption machine

250 people have leveraged a net profit of tens of billions of yuan per year.

Computing power is king: 25,000 GPUs, earning 170 million yuan per day

So how on earth does XTX manage to earn 170 million yuan a day? The answer lies in computing power.

XTX has built a huge GPU cluster to drive its AI trading system.

According to disclosures, XTX owns one of the largest private GPU clusters in the financial trading sector, including more than 25,000 GPUs (10,000 of which are A100 and 10,000 are V100), as well as up to 650PB of available ultra-high-speed storage!

Every day, this "super monster" processes more than one trillion data points!

From tiny fluctuations in global order books to various tick data and sentiment signals, all are processed by the deep learning model.

What XTX stores is the massive amount of transaction data left by the global financial market over decades. Every transaction, every price fluctuation. And the fleeting price linkages between foreign exchange, gold, crude oil, US bonds and stock indexes.

These seemingly messy numbers in the eyes of humans are huge gold mines in the eyes of AI.

Traditional traders look at the market and see greed, fear and the game of human nature. XTX's AI looks at the market and sees a constantly changing high-dimensional probability world.

Every tiny signal may hide a profit opportunity, and the GPU cluster will dig out these signals from massive noise, then execute buy and sell in a fraction of a millisecond.

XTX's daily trading volume in the stock, foreign exchange, fixed income, commodity and cryptocurrency markets exceeds an astonishing 250 billion US dollars!

To support the huge demand for computing power, XTX made an amazing decision: instead of renting cloud servers, build its own data center!

Different from many financial institutions that rely on public clouds, XTX is well aware of the problems of "queuing" and resource preemption in cloud services. In the financial market where every second counts, any delay is fatal.

Therefore, XTX spent more than 1 billion euros to build a 478-acre super data center park in Kajaani, Finland, where the climate is cold and electricity is cheap. The first completed 15,000-square-meter data center is designed to carry up to 22.5 megawatts of computing power.

In this way, XTX can carry out accurate model backtesting and hyperparameter optimization without being constrained by anyone.

Computing power is wealth, and data is the moat.

Reinvest the money earned by AI into the sea of stars of human mathematics

What will Gerko do after earning money that he can't spend for several lifetimes?

This mathematics PhD hopes to give back to mathematical research.

Gerko is extremely passionate about mathematics and geek culture. His investment institution is called "Cromulon Capital", named after an alien god race in *Rick and Morty*.

In the past three years, Gerko and XTX Markets have promised to invest more than 50 million US dollars in open source projects in the field of mathematical AI.

In early 2024, XTX launched the sensational "AIMO Prize", offering a 10 million US dollar reward for the world's first AI model that can win a gold medal in the IMO.

However, the evolution speed of AI far exceeded expectations!

By 2026, XTX said it decided to terminate the AIMO Grand Prize. The reason is: the capability of cutting-edge AI models is improving so fast! From overturning the unit distance conjecture, to automatically formalizing "Fermat's Last Theorem", to open source models easily getting full marks on Olympiad math problems, AI has entered the deep water zone.

Gerko then announced a series of heavy funding plans: investing 10 million US dollars in Lean FRO; another 10 million US dollars in the Mathlib Initiative; and funding SAIR founded by Terence Tao and others.

Gerko, who graduated from Moscow State University with a doctorate, is still a "mathematical fundamentalist" in his bones.

He regards the financial market as an ATM and withdraws hundreds of billions of funds from it.

Then turn these wealth into fuel to push AI to constantly test the boundaries of human logic.

References:

https://www.wsj.com/finance/alex-gerko-xtx-markets-ai-d155626a?st=vBH5VZ&reflink=article_copyURL_share

This article is from the WeChat official account "AI Era", author: ASI Revelation, editor: David, released with authorization from 36Kr.