AI breakthrough from China raises questions about technology access and data security
A Chinese artificial intelligence company is drawing increased attention from international regulators and political leaders as concerns rise over the pace of technological development and global competition in the AI sector.
DeepSeek, a Beijing-based startup, has become a focal point in the ongoing debate over AI governance and national security. The company’s unexpected debut of DeepSeek-V3 earlier this year — a high-performing open-source chatbot trained at a reported cost of just $6 million — has raised eyebrows in Washington and beyond, as governments question how the firm achieved such results at a fraction of the cost faced by Western developers.
US lawmakers consider tighter controls as chip sourcing comes under scrutiny
Sources familiar with internal discussions report that the Trump administration is evaluating a potential crackdown on DeepSeek, including denying the company access to U.S. technology and possibly restricting its platforms within American borders.
A recent report by the U.S. House Select Committee on the Chinese Communist Party suggested DeepSeek may have relied on U.S. technologies and datasets to build its models. It alleges the firm used material derived from American AI models without authorization, and raised serious data privacy concerns by claiming user data could have been routed to Chinese authorities.
Of particular concern is the company’s reported use of approximately 60,000 Nvidia chips in training its models, including an estimated 20,000 units that fall under U.S. export controls. Lawmakers are now investigating how such a volume of restricted hardware may have reached Chinese developers, and whether loopholes or third-party exports played a role.
Rising tensions over AI exports mirror broader US-China trade rift
The growing scrutiny of DeepSeek comes amid broader economic tensions between the United States and China. Washington recently introduced new tariffs of up to 145 percent on a range of Chinese goods, alongside tighter restrictions on the export of semiconductors and advanced AI hardware. Beijing has responded in kind, raising tariffs on some U.S. imports to as much as 125 percent.
Adding fuel to the fire, Singaporean authorities arrested three individuals in February over allegations they exported restricted Nvidia chips — a move local media linked to efforts to supply DeepSeek. The U.S. committee overseeing the investigation has requested detailed records from Nvidia regarding high-volume chip sales in 11 Asian countries since 2020, including Malaysia and Singapore, and any related communication with DeepSeek.
Nvidia has stated that it follows all export rules and that its sales are compliant with U.S. government policies. The company clarified that billing addresses in Singapore may reflect U.S.-based clients with regional offices, and that hardware may ultimately be shipped to countries such as the United States or Taiwan, not mainland China.
Global regulators react to DeepSeek’s rapid expansion
Beyond the United States, regulators in Europe and Asia have launched their own investigations into DeepSeek’s operations. Italy’s data protection agency became the first to block the company’s chatbot over concerns that it failed to meet European data privacy standards. Similar inquiries have been initiated in France, South Korea, and Ireland.
As governments around the world confront the challenge of regulating artificial intelligence, DeepSeek’s rise illustrates both the promise and the peril of rapid innovation. The company has yet to respond publicly to recent allegations and investigations, but its role in the evolving landscape of global AI development is already a matter of significant international consequence.