NVIDIA's shares fell on Monday, as the market value was lost by $ 600 billion-the largest loss of the maximum one-day market in history-where global interest turned to the start of Chinese operation Deepseek.

Technology giant stock Today trading began at $ 140, but it decreased by 16 % and closed at $ 118 as news about Deepseek developments of low -cost artificial intelligence and high -performance spread.

Its dive reflects a wider direction, as it takes off the American technology industry 1 trillion dollars In the market value and those concerns that spread to other areas of the economy, including encryption.

Dibsic capacity It excels Openai's O1 while working at a much lower cost, which is reported less efficiency.

Its success has pushed Deepseek to become the best free application in the United States, according to applications. Data.

Another factor that nourishes the sales and claims that Chinese developers are training artificial intelligence models on NVIDIA H100 chips that prevented the United States NVIDIA from selling, raising questions about the effectiveness of American export controls and China's arrival to advanced devices.

“Chinese laboratories, they have more than 100 people, as you know,“ Scale Scale Alexandr Wang He said CNBC. "What I understand is that Deepseek has about 5000 H100, which they cannot talk about, clearly, because it is against the export controls set by the United States."

The United States ’reluctance to embrace AI Open Source AI will have a advantage in developing artificial intelligence, according to Professor Ion Stoika, the computer world at the University of California at Berkeley and a co -founder of Databricks and Issi, citing organizational concerns and national security concerns.

"When I say the open source, I mean open data, open training algorithms, open weights, and open evaluation - vision on how to train them and what he trained," told Stoika. " Decipher. "We are now in a position in which Chinese companies are betting on open sources, unlike the United States, and it is clear that they are in the future."

While Stoica refused to speculate on the technical stock market, he stressed the transformative capabilities to reduce the costs of the artificial intelligence model.

"If the cost of building or serving these models decreases by 10 or 100x, these companies may harm," he said. "On the other hand, if he pays innovation and speeds up the development of artificial intelligence by making more effort with the same devices, these companies may become more valuable.

Edit Sebastian Senkler