US Accuses Chinese AI Firm DeepSeek of Helping China’s Military and Breaking Chip Rules
— Simplified explanation with a unique angle, based on a detailed report from Reuters
A top U.S. official has revealed that Chinese AI company DeepSeek is helping China’s military and intelligence services, and may be secretly trying to get banned U.S. computer chips, according to a Reuters report.
DeepSeek, based in Hangzhou, shocked the tech world earlier this year by saying its AI models were just as powerful—if not better—than those made by American companies like OpenAI and Meta, but at a much cheaper cost. However, Reuters has reported that U.S. officials now believe the company may be exaggerating its success and relying heavily on technology from U.S. companies like Nvidia.
The U.S. believes DeepSeek is sharing user data with China’s government, even outside the rules of Chinese law, which already allows Beijing to request company data when needed. According to Reuters, the company’s technology has appeared over 150 times in documents related to China’s military, especially its People’s Liberation Army (PLA). Officials say DeepSeek provides tools and services to military research centers in China.
Also worrying U.S. authorities is that DeepSeek has been trying to sneak around U.S. export rules by using fake companies (called shell companies) in Southeast Asia. These fake companies reportedly helped DeepSeek get high-end Nvidia chips (like the powerful H100), which are banned from being shipped directly to China. These chips are used to train advanced artificial intelligence and are crucial for developing powerful technologies—some of which could be used for military purposes. Reuters also shared that DeepSeek may be accessing these chips from outside China using foreign data centers.
Though the U.S. hasn’t officially punished DeepSeek yet or placed it on any trade ban list, a State Department official told Reuters that this kind of behavior is being closely watched. They said the U.S. hasn’t decided yet if more action, like sanctions or new export rules, will be taken against the company.
Meanwhile, Reuters reported that Singapore recently charged three men with fraud in a case connected to Nvidia chips possibly making their way to DeepSeek. And Malaysia is currently investigating whether a Chinese company there is using Nvidia chips to train big AI models.
DeepSeek has not responded to questions from Reuters about the chip issue or their data-sharing practices. However, Nvidia told Reuters that it did not sell DeepSeek the banned chips and that it has stopped doing business with China’s data center market due to U.S. laws. They said DeepSeek likely used a different product, the H800 chip, which was still allowed at the time.
DeepSeek had previously claimed to have tens of thousands of H100 chips, but Reuters sources say the actual number is likely much smaller. Experts are also questioning how DeepSeek could create such advanced AI for just $5.58 million—suggesting the company might be overstating its success.
Unique Perspective: Why This Matters to the World
This isn’t just a tech or political story—it’s about who controls the future of artificial intelligence, and whether that technology will be used to build safer societies or weapons. As Reuters pointed out, China’s growing AI power—possibly fueled by U.S. chips and hidden networks—could give its military an edge, and that’s worrying governments globally.
If DeepSeek is truly feeding data to China’s military and avoiding U.S. export rules, then ordinary people’s privacy and even national security could be at risk—especially since millions of people worldwide use AI tools daily without knowing who is behind them.
With Reuters uncovering these serious claims, the world is now watching not just what DeepSeek does next—but how the U.S. and China respond to growing tensions in the AI and tech race.
Story based on extensive reporting by Reuters.