The United States has ordered government agencies to stop using artificial intelligence tools developed by Chinese startup DeepSeek, citing concerns over the transfer of sensitive data to China, intensifying scrutiny of a company that has disrupted the global AI industry with its low-cost technology.
The restrictions mark a fresh escalation in Washington’s concerns about Chinese AI technology, following earlier bans and limitations on DeepSeek in several countries over data privacy and censorship concerns.
Here are five things to know about the Chinese AI startup, reports AFP.
1. DeepSeek’s ‘Sputnik Moment’ Shook the AI Industry
Founded in 2023 by Liang Wenfeng in Hangzhou, an eastern Chinese technology hub, DeepSeek began as an offshoot of Liang’s data-driven hedge fund, which had access to a stockpile of powerful artificial intelligence processors manufactured by US chipmaker Nvidia.
The company gained global attention in January 2025 with the release of its R1 reasoning model, which triggered a sell-off in US technology stocks and raised questions about the enormous sums American companies were investing in AI development.
Industry insiders were surprised by R1’s reported performance, which rivalled that of leading US chatbots, including ChatGPT, despite DeepSeek’s claims that it had developed the model at a fraction of the cost.
Venture capitalist Marc Andreessen described the breakthrough as a “Sputnik moment”, drawing a comparison with the Soviet Union’s 1957 launch of the first artificial satellite, which shocked the Western world and intensified the space race.
DeepSeek’s emergence challenged assumptions about the cost of developing advanced AI and the technological advantages enjoyed by American companies.
2. Censorship and Data Privacy Concerns
Like other Chinese AI chatbots, DeepSeek’s tools avoid or restrict discussions of politically sensitive subjects that are commonly censored in China, including the 1989 Tiananmen Square crackdown.
Concerns about censorship and the handling of users’ data have prompted several countries to ban or restrict DeepSeek on government-issued devices, including the United States, Australia and South Korea.
On Monday, Washington extended its restrictions by ordering civilian government agencies to stop using DeepSeek products and third-party tools built using its technology.
China’s foreign ministry opposed the move, accusing the United States of “overstretching” the concept of national security.
Despite the concerns, DeepSeek has gained popularity in developing countries, with analysts attributing its appeal partly to its low cost and ease of deployment.
The company accounts for approximately four per cent of the global chatbot market, according to web traffic analysis firm Similarweb, while ChatGPT commands about 68 per cent.
3. Its Open-Source Technology Gives Developers More Freedom
DeepSeek’s models are released with publicly available components that allow developers to inspect, modify and adapt them for different applications, although the extent of access varies by model and release.
This approach is also pursued by major Chinese AI developers, including technology giant Alibaba, and contrasts with the more closed models offered by OpenAI and some other Western competitors.
China has promoted open-source AI as a means of accelerating innovation, expanding adoption and strengthening its position in the global technology industry.
“Chinese AI models are leading the way in the open-source innovation ecosystem,” National People’s Congress spokesman Lou Qinjian told policymakers in April.
The availability of DeepSeek’s technology has helped developers and organisations experiment with AI without necessarily relying on expensive proprietary systems.
However, publicly available model components do not automatically eliminate concerns about data privacy, security or censorship.
4. DeepSeek’s Success Has Boosted China’s AI Industry
DeepSeek’s rise has energised China’s artificial intelligence sector, attracting investor interest despite intensifying competition with the United States and concerns about a possible global technology market bubble.
Shares in Chinese AI startups Zhipu AI and MiniMax surged during their Hong Kong market debuts this year, while Chinese chipmakers, including MetaX, have also attracted investor attention.
In June, investors valued DeepSeek at more than $50 billion during the company’s first funding round, according to the Wall Street Journal.
Shi Yaqiong and her team at Beijing-based Jinqiu Capital told AFP that DeepSeek’s breakthrough had triggered a “clear surge” in enthusiasm for Chinese AI and intensified competition among investors seeking opportunities in the sector.
The company’s progress has also strengthened China’s efforts to develop domestic alternatives to American AI technologies and semiconductor products.
5. Chip Smuggling Allegations Cloud Its Rise
DeepSeek’s progress has also attracted scrutiny over how it obtains the computing power required to develop advanced AI models amid US restrictions on high-end semiconductor exports to China.
Reports, including one by technology publication The Information, alleged that DeepSeek circumvented US export restrictions while training its V4 model.
The Information reported in December, citing six people familiar with the matter, that DeepSeek had developed V4 using thousands of chips dismantled in third countries and smuggled into China.
DeepSeek did not respond to a request for comment at the time.
Nvidia also did not respond directly to a request for comment but told The Information it had seen no evidence supporting the allegations, adding that “such smuggling seems far-fetched”.
In April, DeepSeek said it had used both Nvidia processors and Huawei’s domestically manufactured Ascend chips for its new model.
Huawei, which is at the forefront of China’s efforts to reduce its dependence on American semiconductor technology, said the full range of its Ascend products supports DeepSeek’s V4 series.
Washington has maintained that DeepSeek’s latest model trails the leading offerings from US AI companies by approximately eight months.
The dispute highlights the growing intersection of artificial intelligence, data security, semiconductor export controls and strategic competition between the world’s two largest economies.
