US President Donald Trump has called the erection of the Chinese company Deepseek a “awakening” call to the American technology industry following the appearance of its artificial intelligence model (it) caused shocking waves in Wall Street.
Shares in key technology firms like Nvidia fell sharply, with the chip giant losing nearly $ 600BN (482bn £) in market value.
What has shocked the industry is Deepseek’s claim that its Model R1 is made in part of its rivals’ cost – raising questions about the future of America’s dominance and the degree of investment that US firms are planning.
Deepseek has become the most downloaded app in SH.BA just a week after it was launched.
Responding to news on Monday, Trump said the latest developments in the China industry could be “positive” for the SH.BA
“If you could do it cheaper if you could do it [for] less [and] Go to the same final result. I think this is a good thing for us, ”he told reporters on board Air Force One.
He also said he was not worried about progress, adding that the US will remain a prevailing player on the field.
On Tuesday, shares in Japanese firms linked to him, including Avantest, Softbank and Tokyo Electron dropped significantly, helping to push the Nikkei 225 standard with about 1%.
Some other markets in Asia are closed for the New Year’s lunar holiday.
Continental China’s financial markets will be closed from today and will reopen on February 5th.
Scholarships in the main manufacturing sites of the Taiwan chips and South Korea are also closed today.
Deepseek is powered by the open -sourced Deepseek -v3 model, which his researchers claim was trained for about $ 6 million (4.2m) – significantly less than the billions of rivals spent. But this claim has been discussed by others in him.
Its appearance comes as the US is limiting the sale of advanced chip technology that it empowers in China.
To continue their work without sustainable supplies of imported advanced chips, Chinese developers have shared their work with each other and experimented with new access to technology.
This has resulted in models that require much less computing power than before.
It also means that they cost much less than previously thought possible, which has the potential to raise the industry.
“Deepseek’s ability to rival American models, despite the limited access to advanced equipment demonstrates that software ingenuity and data efficiency can compensate for hardware limitations,” says Marina Zhang, an associate professor at Sydney Technology University, which focuses on China’s high -tech high -tech industries, China’s high -tech, the BBC told.
After the Deepseek-R1 was launched earlier this month, the company boasted “performance at the same time” one of Openai’s last models when used for tasks such as mathematics, coding and natural reasoning of language.
Deepseek’s technology is rated by high profile figures, including Openai chief Sam Altman who called it “an impressive model, especially about what they are able to give for the price”, though he added that Openai would ” obviously provided much better models “moving” moving forward
The Chinese company claims its model can be trained in 2,000 specialized chips compared to about 16,000 for the main models.
But not everyone is convinced. Some have cast doubt on some of Deepseek’s claims, including Elon Musk technology mogul.
He responded to a post that claimed that Deepseek actually has about 50,000 Nvidia chips that have now been banned by the expert in China, saying, “Undoubtedly.”
The company was founded in 2023 by Liang Wenfeng in Hangzhou, a city in Southeast China.
The 40-year-old, a graduate of information and electronic engineering, also established the defense fund that Deepseek supported.
He was recently seen in a meeting between industry experts and Chinese Prime Minister Li Qiang.
In an interview in July 2024 with Academy of ChinaMr. Liang said he was surprised by the reaction to the previous version of his model.
“We didn’t expect prices to be such a sensitive issue,” he said.
“We were simply following our pace, calculating costs and setting prices accordingly.”