The Trump administration is reportedly considering sweeping new restrictions on DeepSeek, a fast-rising Chinese artificial intelligence firm, as part of its broader campaign to slow China’s advance in next-generation technologies. The moves, first reported by The New York Times, could significantly reshape the global AI landscape and deepen the ongoing US-China tech rivalry.
DeepSeek’s Disruptive Rise
Founded by Liang Wenfeng, DeepSeek gained prominence earlier this year after launching its open-source model DeepSeek-V3, which reportedly cost just $6 million to develop. The model’s affordability and performance have attracted global attention, posing a direct challenge to US AI leaders like OpenAI. DeepSeek also unveiled a new version of its chatbot “R1” in January, claiming superior performance to ChatGPT despite using outdated microchips.
Read More: DeepSeek: China’s AI Game-Changer
The company’s rapid ascent has triggered alarm in Washington, where officials fear that such low-cost, high-performing models could undercut US dominance in AI and lead to the misuse of advanced technology.
Potential Penalties Under Review
US officials are now weighing penalties that could prevent DeepSeek from purchasing American technology—particularly advanced AI chips from Nvidia—and may even bar American users from accessing DeepSeek’s services. These measures are part of a broader strategy to limit China’s access to cutting-edge AI hardware and services.
The Trump administration this week further tightened export restrictions on Nvidia, specifically targeting its H20 chip, which had been tailored for the Chinese market. Nvidia warned it could lose up to $5.5 billion in revenue due to these restrictions.
Congressional Concerns and Security Threats
A recent report by the bipartisan US House Select Committee on the Chinese Communist Party declared DeepSeek a “profound threat” to US national security. The committee raised several concerns:
- Data Harvesting: Allegations that DeepSeek collects data from American users and transmits it to Chinese authorities.
- Military and State Ties: Claims that DeepSeek operates within a tightly integrated ecosystem involving Chinese state-linked research institutions and hardware providers.
- IP Theft: OpenAI has accused DeepSeek of distilling its proprietary models in violation of its terms of service. Lawmakers also suspect that DeepSeek may have trained its systems using unlawfully obtained US intellectual property.
Nvidia in the Crosshairs
The congressional committee has demanded answers from Nvidia regarding how its chips—specifically those covered under US export controls—may have ended up powering DeepSeek’s AI models. Some estimates suggest DeepSeek has acquired up to 60,000 Nvidia chips, including 20,000 that should have been restricted.
Singaporean authorities recently arrested three individuals for allegedly smuggling Nvidia chips to DeepSeek, further intensifying scrutiny over export violations. Nvidia maintains that it adheres strictly to US laws and export controls.
Global Scrutiny and Export Control Challenges
Beyond the US, DeepSeek is also under investigation in multiple countries, including Italy, France, South Korea, and Ireland, for potential violations of data privacy regulations. The broader backlash reflects growing international unease over China’s AI ambitions and regulatory gray areas surrounding AI deployment.
Despite these efforts, critics argue that previous US export restrictions may have backfired, accelerating China’s efforts to develop indigenous chip technology and alternative AI solutions. This raises questions about the long-term efficacy of containment strategies.
Read More: Best Countries to Visit in Middle East in 2024
The DeepSeek controversy marks a major escalation in the global AI tech war. With Washington considering investment restrictions, tighter export rules, and potential service bans, the US is doubling down on efforts to preserve its technological edge. Meanwhile, China continues to make gains in foundational AI, signaling a prolonged and complex contest over the future of artificial intelligence.