The 500 richest people in the world, pulled by Nvidia co-founder Jensen Huang, lost a total combined of $ 108 billion on Monday (27), as a technology-led sales-led sales of Deepseek developer Deepseek ,
Billionaires whose fortunes are linked to artificial intelligence were the biggest losers: Huang saw his fortune fall $ 20.1 billion, a 20%drop, while the loss of $ 22.6 billion from Oracle Cofunder Larry Ellison, It was higher in absolute terms, but it represented only 12% of its fortune, according to Bloomberg’s billionaires. Dell’s Michael Dell lost $ 13 billion, and Changpeng “CZ” Zhao, co -founder of Binance, saw his fortune decrease by $ 12.1 billion.
Titans in the technology industry, as a whole, saw $ 94 billion of richness evaporate – about 85% of the total fall of the Bloomberg index. The Nasdaq Composite index fell 3.1%, and the S&P 500 fell 1.5%.
Deepseek, based in Hangzhou, has been developing AI models since 2023, but the company has caught the attention of many western investors this weekend, when its free chatbot application Deepseek R1 led downloads worldwide. So many new users have signed up that DeepSeek had difficulty keeping the app online, facing falls and forcing it to restrict registration to users with Chinese telephone numbers.
Deepseek’s unexpected entry into the AI race, which according to the company cost only $ 5.6 million to be developed, challenges silicon valley narrative that huge capital spending is essential to developing the strongest models. This was a hard blow to billionaires whose fortunes are linked to the Western Ia supply chain, which has been the largest engine market engine in the last two years.
Similar script
High reviews for the so -called hyperscalers de AI – including Meta, Alphabet and Microsoft – have generated billions in wealth for their owners since OpenAi launched ChatgPT in November 2022. These companies, for the most part, operated with a similar script: spending huge amounts to develop and operate systems AI, accumulating state-of-the-art semiconductors and the energy supplies needed to operate them.
Continues after advertising
The CEO of the goal, Mark Zuckerberg, announced on Friday that the company planned to spend from $ 60 billion to $ 65 billion on AI-related projects this year, well above Wall Street estimates. Capital spending on all major technology companies are on their way to reaching $ 200 billion in 2025, according to a Bloomberg Intelligence report.
Despite the limited revenue to justify all their investments so far, markets have rewarded US technology actions with record evaluations, which in turn generated historical wealth gains for their owners. Nvidia has stood out as the largest individual AI boom winner so far, with Huang’s fortune increasing almost eight times to $ 121 billion from the beginning of 2023 until Friday. Zuckerberg’s fortune rose 385% to $ 229 billion in the same period and Jeff Bezos, from Amazon.com., increased by 133% to $ 254 billion.
Although Huang and Ellison suffered losses, the fortunes of other major technology billionaires came unharmed. Zuckerberg’s fortune ended the day, earning $ 4.3 billion as the goal recovered from a fall at the beginning of the session. Bezos’ richness rose about $ 632 million.
Continues after advertising
Capital expenses
The fact that Deepseek has been able to develop a free model that potentially rivals or surpasses competitors such as Anthropic’s ChatgPT and Claude to a fraction of the cost of development led investors to question the logic behind Silicon Valley dependence on the silicon capital spending.
A key reason why DeepSeek didn’t depend on large investments and cutting -edge chips to develop its model is that Chinese companies had limited access to powerful GPUs, or graphic processing units, of which most western companies depends since the government from the US has instituted strict export controls on the most advanced chips.
In an interview with CNBC last week, Alexandr Wang, CEO of the Scale Ai training data supplier, said that despite export controls, Deepseek and other Chinese developers probably have more GPUS than Silicon Valley imagines.
Continues after advertising
“Chinese laboratories have more h100s than people think,” Wang said, referring to Nvidia’s toy AI chip. “My understanding is that DeepSeek has about 50,000 h100s, which they can’t speak obviously because it goes against the export controls the US has in force.”
© 2025 Bloomberg L.P.
