DeepSeek: How China’s ‘AI heroes’ overcame US curbs to stun Silicon Valley

Fan Wang and João de Silva

BBC News

Reporting fromSingapore
Getty Images A Chinese woman checks her phone, with a sign in the background reads "I heart Beijing"Getty Images

When ChatGpt stormed the world of artificial intelligence ( AI), an obvious problem followed: did it spell trouble for China, America’s biggest tech enemy?

Two decades on, a fresh AI type from China has flipped that question: does the US quit Chinese development?

For a while, Beijing seemed to struggle with its response to ChatGPT, which is not attainable in China.

Customers who were unresponsive made fun of Baidu’s search engine giant Ernie the robot. Therefore, there were versions from tech companies Tencent and ByteDance, both of which were criticized as being less than good.

Washington wanted to keep things that way and was convinced in its position. So the Biden administration imposed more limits on the trade of cutting-edge technology and products to China.

That’s why DeepSeek’s establish has astonished Silicon Valley and the universe. The business claims that its innovative design is much less expensive than the US$ billion spent on AI.

How did a little-known business, whose leader is being hailed as an” AI hero” on Chinese social media, pull this off?

The problem

It was undoubtedly a blow when the US prohibited the country’s top chip manufacturers from selling cutting-edge technology to China.

These cards are necessary for creating potent AI models capable of performing a variety of individual tasks, from simple queries to complex math problems.

DeepSeek’s leader Liang Wenfeng described the device ban as their “main problem” in conversations with native media.

Much before the restrictions, DeepSeek acquired a” large stock” of Nvidia A100 cards- estimates range from 10, 000 to 50, 000- according to the MIT Technology Review.

Leading AI types in the West employ an approximated 16, 000 professional cards. But DeepSeek says it trained its Artificial type using 2, 000 like cards, and hundreds of lower-grade chips- which is what makes its goods cheaper.

Elon Musk, a US tech tycoon, and other figures have disputed this assertion, claiming that the company never show how many advanced chips it actually used given the limitations.

But authorities say Washington’s restrictions brought both challenges and opportunities to the Chinese AI market.

According to Marina Zhang, an associate professor at the University of Technology Sydney, it has “forced Taiwanese firms like DeepSeek to develop” so they can do more with less.

CCTV A photo showing Liang Wenfung attending a meeting chaired by Chinese Premier Li Qiang on 20 January.CCTV

” While these limitations pose problems, they have also spurred creativity and endurance, coordinating with China’s broader plan goals of achieving technical independence”.

From the batteries that power electric cars and solar panels to AI, the second-largest economy in the world has invested a lot in great technology.

Turning China into a technical power has long been President Xi Jinping’s passion, so Washington’s regulations were also a problem that Beijing took on.

The transfer of DeepSeek’s new design on 20 January, when Donald Trump was sworn in as US senator, was intentional, according to Gregory C Allen, an AI specialist at the Center for Strategic and International Studies.

According to Mr. Allen, former chairman of approach and plan at the US Department of Defense Joint Artificial Intelligence Center,” the timing and the way it’s being messaged is precisely what the Taiwanese government wants everybody to believe,” he said.” The timing and the way it’s being messaged is precisely what the Chinese government wants everyone to believe.

In recent years, the Chinese government has encouraged collaboration between universities and industry by providing scholarships and research grants.

According to Ms. Zhang, the National Engineering Laboratory for Deep Learning and other state-backed initiatives have assisted in the training of thousands of AI specialists.

Additionally, China had plenty of talented engineers to hire.

The talent

Take DeepSeek’s team for instance- Chinese media says it comprises fewer than 140 people, most of whom are what the internet has proudly declared as “home-grown talent” from elite Chinese universities.

Western observers disregarded the “new generation of entrepreneurs who prioritize foundational research and long-term technological advancement over quick profits,” according to Ms Zhang.

Even the youngest managers are frequently under the age of 35, thanks to China’s top universities’ rapidly expanding AI talent pool.

” Having grown up during China’s rapid technological ascent, they are deeply motivated by a drive for self-reliance in innovation”, she adds.

Getty Images An aerial view shows the Alibaba Digital Eco-Innovation Park in Hangzhou, ChinaGetty Images

Deepseek’s founder Liang Wenfeng is an example of this- the 40-year-old studied AI at the prestigious Zhejiang University. People with whom he is known claim in an article on the tech outlet 36Kr that he is “more like a geek than a boss.”

And Chinese media describe him as a “technical idealist”- he insists on keeping DeepSeek as an open-source platform. In fact, experts also think that young start-ups have gained more from a thriving open-source culture and faster growth as a result.

Unlike bigger Chinese tech firms, DeepSeek prioritised research, which has allowed for more experimenting, according to experts and people who worked at the company.

In an interview with 36Kr, Mr. Liang said,” The Top 50 talents in this field might not be in China, but we can build people like that here.”

However, experts are unsure of how far DeepSeek can advance. According to Ms. Zhang, “new US restrictions may limit access to American user data, potentially affecting how Chinese models like DeepSeek can expand internationally.”

And others say the US still has a huge advantage, such as, in Mr Allen’s words,” their enormous quantity of computing resources”- and it’s also unclear how DeepSeek will continue using advanced chips to keep improving the model.

Given that the majority of Chinese people had never heard of it up until this weekend, DeepSeek is enjoying its moment in the sun for the time being.

The new AI heroes

Mr. Liang’s sudden fame has caused him to gain notoriety on China’s social media platforms, where he is being hailed as one of the” three AI heroes” from the southern Guangdong province, which borders Hong Kong.

The other two are Zhilin Yang, a leading expert at Tsinghua University, and Kaiming He, who teaches at MIT in the US.

DeepSeek has delighted the Chinese internet ahead of Lunar New Year, the country’s biggest holiday. Good news for a struggling economy and a tech sector that is anticipating additional tariffs and the potential sale of TikTok’s US business.

A popular Weibo comment reads,” DeepSeek shows us that only the real deal will stand the test of time.”

” This is the best new year gift. Wish our motherland prosperous and strong”, another reads.

A “blend of shock and excitement, particularly within the open-source community”, is how Wei Sun, principal AI analyst at Counterpoint Research, described the reaction in China.

Getty Images Visitors enjoy illuminated red lanterns to celebrate the Spring Festival in ChinaGetty Images

Fiona Zhou, a tech worker in the southern city of Shenzhen, says her social media feed “was suddenly flooded with DeepSeek-related posts yesterday”.

” People call it’ the glory of made-in-China’, and say it shocked Silicon Valley, so I downloaded it to see how good it is”.

She asked it for “four pillars of]her ] destiny”, or ba-zi- like a personalised horoscope that is based on the date and time of birth.

But to her disappointment, Deepseek was wrong. While she was given a thorough explanation about its” thinking process”, it was not the “four pillars” from her real ba-zi.

She claims she will continue to try it because it will likely be more useful for such tasks.