With the release of an artificial intelligence (AI) model that rivals the capabilities of Google and OpenAI, DeepSeek, a little-known Chinese startup, has shocked the world’s tech industry.
DeepSeek-R1’s creator says its model was developed using less advanced, and fewer, computer chips than those employed by tech giants in the United States.
The model’s development team claimed to have spent less than $6 million on computing power to train the model, a fraction of the multibillion-dollar AI budgets enjoyed by US tech giants like OpenAI, Alphabet, and Meta, in a research paper released last week.
Marc Andreessen, one of the most influential tech venture capitalists in Silicon Valley, hailed the release of the model as “AI’s Sputnik moment”.
The recent emergence of a small Chinese startup capable of outperforming Silicon Valley’s top players has questioned assumptions about US supremacy in AI and raised concerns that the sky-high market valuations of companies like Nvidia, Alphabet, and Meta may be untrue.
On Monday, Nvidia, which holds a near-monopoly on producing the semiconductors that power generative AI, lost nearly $600bn in market capitalisation after its shares plummeted 17 percent.
US President Donald Trump, who last week announced the launch of a $500bn AI initiative led by OpenAI, Texas-based Oracle and Japan’s SoftBank, said DeepSeek should serve as a “wake-up call” on the need for US industry to be “laser-focused on competing to win”.
What is DeepSeek?
DeepSeek, which is based in Hangzhou, was founded in late 2023 by Liang Wenfeng, a serial entrepreneur who also runs the hedge fund High-Flyer.
Although little known outside of China, Liang has a long history of combining investing with cutting-edge technology.
In 2013, he co-founded Hangzhou Jacobi Investment Management, an investment firm that employed AI to implement trading strategies, along with a co-alumnus of Zhejiang University, according to Chinese media outlet Sina Finance.
In 2015 and 2016, Liang established two additional companies with a focus on computer-directed investing, Hangzhou Huanfang Technology Co. and Ningbo Huanfang Quantitative Investment Management Partnership, respectively.
In an interview with the Chinese media outlet Waves in 2023, Liang refuted the claim that startups should be too late to start using AI or that it was too expensive.
“Reproduction alone is relatively cheap — based on public papers and open-source code, minimal times of training, or even fine-tuning, suffices. Research, however, involves extensive experiments, comparisons, and higher computational and talent demands”, Liang said, according to a translation of his comments published by the ChinaTalk Substack.
Liang claimed that curiosity was the main factor in his interest in AI.
“From a broader perspective, we want to validate certain hypotheses. For example, we hypothesise that the essence of human intelligence might be language, and human thought could essentially be a linguistic process”, he said, according to the transcript.
What you perceive as “thinking” might actually be your brain-waving language. This suggests that large language models with AGI-like capabilities could develop, he added, referring to artificial general intelligence (AGI), a branch of artificial intelligence that attempts to imitate the cognitive abilities of people.
A comment request was not immediately responded to by DeepSeek.
On Monday, Gregory Zuckerman, a journalist with The Wall Street Journal, said he had learned that Liang, who he had not heard of previously, wrote the preface for the Chinese edition of a book he authored about the late American hedge fund manager Jim Simons.
In a column describing Liang’s book as a tome that “unravels many previously unresolved mysteries and brings us a wealth of experiences to learn from,” Zuckerman wrote, “Simons left a deep impact, apparently.”
“Even my mother didn’t get that much out of the book”, Zuckerman wrote.
Why has DeepSeek completely transformed the tech industry?
Simply put, Silicon Valley and the US government’s approach to AI have been questioned by the success of the company.
Due to their size, which allows them to attract the best talent from all over the world and invest colossal amounts in building data centers and purchasing large quantities of expensive high-end chips, are widely believed to have a significant edge in AI.
The arrival of DeepSeek has challenged the notion that a billion-dollar AI startup must be in the top echelon.
“OpenAI was founded 10 years ago, has 4, 500 employees, and has raised $6.6 billion in capital. DeepSeek was founded less than 2 years ago, has 200 employees, and was developed for less than $10 million”, Adam Kobeissi, the founder of market analysis newsletter The Kobeissi Letter, said on X on Monday.
“How are these two companies now competitors”?
In their research paper, DeepSeek’s engineers said they had used about 2, 000 Nvidia H800 chips, which are less advanced than the most cutting-edge chips, to train its model.
The team claimed to have used a number of highly sophisticated models to combine to make slower chips’ data analyses more effective.
For the US government, DeepSeek’s arrival on the scene has raised questions about its strategy of trying to contain China’s AI advances by restricting exports of high-end chips.
According to the research paper from DeepSeek, either the most sophisticated chips are required to produce high-performing AI models or Chinese companies can still obtain the chips in sufficient quantities, or a combination of both.
California-based Nvidia’s H800 chips, which were designed to comply with US export controls, were freely exported to China until October 2023, when the administration of then-President Joe Biden added them to its list of restricted items.
In his 2023 interview with Waves, Liang said his company had stockpiled 10, 000 Nvidia A100 GPUs before they were banned for export. Graphic processing units, or graphics processing units, are electronic circuits that are used to speed up computer graphics and image processing.
Given the success of various AI models being developed by Chinese companies like Alibaba and Baichuan, Tanishq Abraham, a former research director at Stability AI, said he was not surprised by China’s level of AI advancement.
China has been able to innovate and squeeze performance from whatever they have, Abraham said to Al Jazeera, “even though there have been restrictions on China’s ability to obtain GPUs.”
“I believe that US companies should learn that there is still a lot of performance they can squeeze out of.”
Tara Javidi, co-director of the Center for Machine Intelligence, Computing and Security at the University of California San Diego, said DeepSeek made her excited about the “rapid progress” taking place in AI development worldwide.
According to Javidi, “my only hope is that the attention given to this announcement will increase the level of intellectual interest in the subject, further increase the talent pool, and, last but not least, increase both private and public funding for AI research in the US,” Javidi said.
Meanwhile, investors ‘ confidence in the US tech scene has taken a hit – at least in the short term.
Apart from Nvidia’s dramatic slide, Google parent Alphabet and Microsoft on Monday saw their stock prices fall 4.03 percent and 2.14 percent, respectively, though Apple and Amazon finished higher.
“If DeepSeek’s cost numbers are real, then now pretty much any large organisation in any company can build on and host it”, Tim Miller, a professor specialising in AI at the University of Queensland, told Al Jazeera.
“So, in this sense, the game has changed completely because there is a new ‘ rule ‘ that anyone can play”.
Does this indicate that China is winning the AI race?
Not necessarily.
The field is moving quickly despite the consensus that ChatGPT and DeepSeek-R1 perform at comparable or even better levels for some tasks.
Sam Altman, the company’s CEO, announced earlier this month that the company would release its most recent reasoning AI model, the o3 mini, within weeks after taking user feedback into consideration.
On Monday, Altman acknowledged that DeepSeek-R1 was “impressive” while defending his company’s focus on greater computing power.
“We will undoubtedly deliver much better models,” and having a new rival is legit energizing! We will pull up some releases”, Altman said on X.
“But most of all, we are eager to carry out our research roadmap and think that more computation is needed now than ever to fulfill our mission.”
Rui Ma, the founder of Tech Buzz China, claimed that the Chinese tech industry was shocked when OpenAI first released ChatGPT in 2022.
“Most entrepreneurs had completely missed the opportunity that generative AI represented, and felt very humbled”, Ma told Al Jazeera.
“It’s clear that they have been hard at work since. This past weekend, in my opinion, demonstrates how seriously they took the challenge of “catching up” with Silicon Valley. Washington should concentrate on boosting Silicon Valley rather than repressing China, in my opinion, to keep the US in the lead.
According to Abraham, the former research director at Stability AI, perceptions may also be influenced by the fact that companies like OpenAI haven’t made their most advanced models freely available to the general public, in contrast to DeepSeek.
The best model available for free use was made by DeepSeek. On the other hand, OpenAI’s best model is not free”, he said.
“So DeepSeek shocked most users who use ChatGPT for free and gives the impression that there are significantly more capabilities now than OpenAI, which has already had a similar-performing model paidwalled for a few months.” People are unable to fully understand the advancement and capabilities of AI due to this pay-walling of frontier AI models.
Miller, the University of Queensland professor, said DeepSeek’s advances and other recent developments suggest that China is at least “up there” with the US in AI.
Late last year, I made a somewhat rash prediction that the next scientific discovery in AI could be made by a small-scale researcher without much computing power, arguing that they would need to be smarter to compete.
Source: Aljazeera
Leave a Reply