How Chinese “Ai Heroes” overcome US curbs to stun the Silicon Valley

When Chatgpt storms the world of artificial intelligence (AI), there was an inevitable question: Did he create problems for China, the largest technological rival in America?

Two years on, a new AI model from China has transferred this question: Can the US stop Chinese innovation?

For a while, Beijing seems to have encountered his response to Chatgpt, which is not available in China.

Unnoticed users made fun of Ernie, The Chatbot from the Baidu Search Engine giant. Then came versions of Tencent and Bytedance technology companies, which were rejected as Chatgpt followers – but not so good.

Washington was confident that he was ahead and wanted to keep it that way. Thus, the Biden administration has increased restrictions by banning the exports of advanced chips and technologies to China.

That is why the launch of Deepseek amazed the Silicon Valley and the World. The company says its powerful model is far more expensive than billions of companies that have spent on AI.

So, how did the little known company – whose founder is welcomed into Chinese social media as a “AI hero” – to take it out?

When the United States banned the world's leading chip manufacturers such as Nvidia from selling Advanced Tech in China, it was certainly a blow.

These chips are essential for the construction of powerful AI models that can perform a number of human tasks, from response to basic requests to solving complex mathematical problems.

Deepseek Founder Liang Venfen described the ban on the chip as their “main challenge” in interviews with local media.

Long before the ban, Deepseek acquired “significant stocks” of the NVIDIA A100 chips – ratings range from 10,000 to 50,000 – according to Mit Technology Review.

The leading AI models in the West use approximately 16,000 specialized chips. But Depepeek says he has trained his AI model using 2000 such chips, and thousands of chips with a smaller class -which makes his product more expensive.

Some, including the American technology billionaire Elon Musk, have asked this statement, arguing that the company cannot reveal how many advanced chips it has really used, given the restrictions.

But experts say the ban on Washington brought both challenges and the opportunities of the Chinese and industry.

He has “forced Chinese companies like Deepseek to innovate,” so they can do more with less, says Marina Gian, an associate professor at Sydney Technical University.

“While these restrictions are challenges, they also triggered creativity and sustainability, giving China's broader political goals to achieve technological independence.”

The world's second largest economy has invested strongly in large technologies – from batteries that power electric vehicles and solar panels to AI.

Turning China into a technological superpower has long been the ambition of its President Jinping, so Washington's restrictions were also a challenge that Beijing has taken.

The release of the new Deepseek model on January 20, when Donald Trump was sworn in as President of the United States, was deliberate, according to Gregory with Alan, an AI expert at the Strategic and International Research Center.

“The time and the way messages are being sent – this is exactly the Chinese government wants everyone to think – that export control does not work and that America is not a world leader in AI,” says G -n -Allen, a former director of strategy and politics in The Joint Center for Artificial Intelligence of the United States.

In recent years, the Chinese government has nourished AI talent by offering scholarships and grants for research and promoting partnerships between universities and industry.

The National Engineering Laboratory for Deep Training and other state initiatives have helped the training of thousands of EIA specialists, according to Mrs. Gian.

And China had very bright dialing engineers.

For example, get the Deepseek team – Chinese media say it consists of less than 140 people, most of whom are what the Internet proudly declared “talent grown at home” by elite Chinese universities.

Western observers have missed the emergence of a “new generation of entrepreneurs who give priority to fundamental research and long-term technological progress against quick profits,” says Mrs. Gian.

The best universities in China create a “fast -growing AI talents”, where even managers are often under 35 years of age.

“After growing up during the rapid technological climb of China, they are deeply motivated by the pursuit of independence in innovation,” she adds.

Deepseek Liang Venfen's founder is an example of this-40-year-old studies AI at the prestigious Zhejiang University. In an article on Tech Outlet 36KR, people familiar with it say he is “more like a manner rather than a boss.”

And the Chinese media describe it as a “technical idealist” – he insists on keeping a Deepseek as an open source platform. In fact, experts also believe that the open-source-thriving culture has allowed young start-ups to combine resources and progress faster.

Unlike larger Chinese technology companies, Depepeek prioritizes research that has allowed more experiments, according to experts and people who have worked at the company.

“The best 50 talents in this area may not be in China, but here we can build such people,” said G -n Liang in an interview with 36KR.

But experts are wondering how much more can be divided. D -Jan says that “new American restrictions can limit access to US consumer data by potentially influencing how Chinese models like Deepseek can go global.”

And others say the United States still have a huge advantage, such as, in the words of G -n Alan, “their huge amount of computing resources” – and it is also not clear how Deepseek will continue to use sophisticated chips to continue to Improves the model.

But so far, Deepseek is enjoying its moment in the sun, given that most people in China have never heard of it until this weekend.

His sudden fame saw that he became a sensation in the social media of China, where he was applauded as one of the “three heroes of the AI” by the province of South Guangdong, who borders Hong Kong.

The other two are Giller Yang, a leading expert at Tsinghua University, and Kaiming, who teaches at MIT in the United States.

Deepseek delighted the Chinese Internet before the Lunar New Year, the biggest vacation in the country. This is good news for a dressed economy and the technology industry, which is preparing for more tariffs and the possible sale of Tiktok US business.

“Deepseek shows us that only if you have a real deal will you withstand the time test,” said Weibo's highest liked comment.

“This is the best New Year's gift. Wish our homeland prosperous and strong,” said another.

“A mixture of shock and excitement, especially in the open code community,” is how Wei Sun, AI chief analyzer at Counterpoint Research, described the reaction in China.

Fiona Zhou, a technology worker in the southern city of Shenzhen, says her issue on social media “was suddenly flooded with posts related to Deepseek.”

“People call it” the glory of what has been done in China “and say she shocked the Silicon Valley, so I pulled it to see how good it was.”

She asked for it for “four pillars of (her) fate” or ba -zi – as a personalized horoscope based on the date and time of birth.

But for her disappointment, Deepseek is wrong. As she received an in-depth explanation for his “thinking process”, it was not the “four pillars” of her real Ba-Zi.

She says she will still give it another job, as it will probably be more useful for such tasks.