The Man Who Wiped Out $600 Billion by Combining Quantitative Finance with Technology

DeepSeek was created by a leading Chinese quantitative trading firm. It’s not uncommon for Wall Street celebrities in quantitative finance to make breakthroughs in technology. Nvidia Corp. shares erased $593 billion in market capitalization on Monday, marking the largest single-day loss in U.S. market history, according to data. One man was primarily responsible. His name is Liang Wenfeng, and he is the co-founder of Chinese quantitative hedge fund High-Flyer Quant and its artificial intelligence business, DeepSeek.

High-Flyer is one of the largest quantitative hedge funds operating in the Chinese markets, according to the South China Morning Post, utilizing algorithms, mathematical models, and computers to trade financial markets.

DeepSeek gained international attention over the weekend after unveiling its latest model. U.S. users can access it via a mobile app. By the end of Sunday, DeepSeek’s app climbed to the top spot among the most downloaded apps in Apple Inc.’s (AAPL) store.

Meanwhile, social media platforms like X were buzzing with reviews that the new DeepSeek model rivals publicly available Western models from OpenAI and Anthropic. After DeepSeek launched its latest model on January 20, some metrics showed it performed at least as well as mainstream products from OpenAI, creators of the ChatGPT large language model family, and Anthropic, creators of Claude.

Moreover, DeepSeek achieved this on a limited budget and – according to the company – without the help of Nvidia’s most advanced chips.

The unexpected success of DeepSeek caused turbulence in U.S. stocks on Monday as investors considered whether the U.S. could lose its status as the global leader in artificial intelligence and whether Nvidia’s expensive chips will always be essential for fueling AI innovation.

The model’s performance also raised uncomfortable questions about whether the vast sums spent by companies like Microsoft Corp. (MSFT), Alphabet Inc. (GOOGL) (GOOG), and Amazon.com Inc. (AMZN) on infrastructure to enhance their AI capabilities will ultimately pay off.

Shares of all three companies declined on Monday, though their losses were not as significant as Nvidia’s. The leading chipmaker for AI saw its stock close 16.9% lower, translating to a market capitalization loss of $588.9 billion, according to Dow Jones Market Data.

Who is Liang?

Liang, 40, grew up in the southeastern Chinese province of Guangdong and attended the prestigious Zhejiang University in Hangzhou, where he studied electronic information and computer programming, according to a report in SCMP. He founded High-Flyer along with two university colleagues in 2015.

Earlier this month, Liang attended a symposium organized by Chinese Premier Li Qiang in Beijing, according to SCMP. His presence was seen as a sign of DeepSeek’s growing significance to China’s emerging AI industry.

High-Flyer announced that it began building Nvidia chip clusters several years ago. DeepSeek claims its models were trained on fewer than 10,000 Nvidia A100 GPUs, a claim that sparked criticism in the U.S., including from Tesla Inc. (TSLA) CEO and GrokAI founder Elon Musk.

Liang’s hedge fund manages assets worth approximately $8 billion, but he is far from the first in quantitative finance to use his skills in trading to create technologies with broader applications.

Years before Liang arrived in Hangzhou to study at its prestigious university, Jeff Bezos, who would later found Amazon.com, worked at D.E. Shaw & Co., a pioneering quantitative trading firm on Wall Street, before leaving to start his own company. Bezos used the skills he gained at D.E. Shaw to create an online bookstore, which grew over time to sell virtually everything to end consumers. Amazon also built a massive cloud platform called Amazon Web Services.

When Quantitative Finance and Technology Intersect

According to Garrett Shepherd, head of Voya Machine Learning Intelligence, there is significant overlap in the skills needed to succeed in both quantitative investing and AI research.

“Even the technology used in quantitative finance is the same set of tools that major tech companies use,” Shepherd said. “Reinforcement learning, deep learning, natural language processing, and foundational networks – all these AI techniques and approaches are applicable in both fields.”

Liang’s decision to expand his focus from finance to AI underscores how competitive – and potentially lucrative – the artificial intelligence field is expected to be.

Over the years, Wall Street has had an edge in attracting top tech talent, said Evan Fiegens, co-manager of TCW Artificial Intelligence ETF AIFD, in an interview.

“For a long time, quantitative hedge funds aggressively hired specialists from the tech sector,” Fiegens said. “There aren’t many people with this kind of expertise in data science and machine learning, and working in a quantitative hedge fund was one of the most lucrative things you could do.”

But with the advent of the AI revolution, some of the big tech companies are again gaining the upper hand. Given the immense value of developing these models, companies like OpenAI are willing to pay whatever it takes to attract top talent.

“Now we’re seeing this talent divide between tech and AI starting to tilt back towards technology,” Fiegens said.