US President Donald Trump called for the rise of Chinese company Deepseek “an alarm clock” for the American technology industry, after the emergence of his artificial intelligence model (IA) triggered shock waves at Wall Street .
The actions of large technological companies like Nvidia have dropped sharply, the chip giant losing more than $ 500 billion (400 billion pounds sterling) in market value.
What rocked the industry was the assertion of Deepseek that his R1 model was made to a fraction of the cost of its competitors – raising questions about the future of the domination of American AI and The investment scale that US companies provide.
Deepseek has become the most downloaded free application in the United States a week after its launch.
Responding to the news on Monday, Trump said that the latest developments in the Chinese AI industry could be “positive” for the United States.
“If you could do it cheaper, if you could do it (for) less (and) reach the same final result. I think it’s a good thing for us,” he told journalists on board Air Force One.
He also said that he was not concerned about the breakthrough, adding that the United States will remain a dominant actor on the ground.
On Tuesday, actions in Japanese companies linked to AI, including Adventst, SoftBank and Tokyo Electron, fell sharply, helping to lower the Nikkei 225 by around 1%.
Several other markets in Asia are closed for the new year lunar holidays.
The financial markets of continental China will be closed from today and will reopen on February 5.
Scholarships in the main producing countries of Taiwan fleas and South Korea are also closed today.
Deepseek is fueled by the open source Deepseek -V3 model, which, according to its researchers, was formed for approximately $ 6 million (4.2 million pounds sterling) – much less than the billions spent by competitors. But this assertion was challenged by others in AI.
Its emergence occurs while the United States restricts the sale of advanced chip technology that feeds AI in China.
To continue their work without stable supplies of imported advanced fleas, the developers of Chinese AI have shared their work between them and have experienced new approaches to technology.
This has led to AI models that require much less computing power than before.
It also means that they cost much less than possible than possible, which has the potential to upset the industry.
“Deepseek’s ability to compete with American models despite limited access to advanced hardware shows that software ingenuity and data efficiency can compensate for hardware constraints,” explains Marina Zhang, Associate Professor at the University of Technology Technology Sydney, which focuses on high -tech Chinese industries, said to the BBC
After Deepseek-R1 was launched earlier this month, the company boasted of “performance on a par with” one of the last models of Openai when used for tasks such as mathematics, coding and reasoning in natural language.
Deepseek technology was praised by high -level figures, including Openai Sam Altman chief who called him “an impressive model, especially around what they can deliver for the price”, although He added that Optai “would obviously provide much better models” in moving before.
The Chinese company claims that its model can be formed in 2,000 specialized chips, against around 16,000 for the leading models.
But not everyone is convinced. Some have questioned some of Deepseek’s claims, including the Elon Musk technology magnate.
He responded to a position which said that Deepseek actually has around 50,000 Nvidia chips which have now been prohibited from experts in China, saying: “Obviously”.
The company was founded in 2023 by Liang Wenfeng in Hangzhou, a city in southeast China.
The 40 -year -old man, a graduate of information and electronic engineering, also founded the Hedge Fund which supported Deepseek.
It was recently seen at a meeting between industry experts and Chinese Prime Minister Li Qiang.
In an interview in July 2024 with China Academy, Mr. Liang said he was surprised by the reaction to the previous version of his AI model.
“We did not expect prices to be such a sensitive problem,” he said.
“We simply followed our own rhythm, let us calculate the costs and fix the prices accordingly.”
