By Maximilian Marweld
China’s rapid advances in artificial intelligence (AI) have challenged the effectiveness of US technological containment policies. The recent emergence of DeepSeek-R1—a Chinese developed AI model comparable to OpenAI’s GPT-4—has demonstrated that China remains highly adaptive despite stringent export controls on high-performance computing chips. Developed on a modest $6 million budget without access to NVIDIA’s most advanced GPUs, DeepSeek illustrates Beijing’s ability to weather heavy restrictions on modern technology. As The Economist aptly put it, DeepSeek’s success “embodies China’s ability to turn constraints into advantages.”
China’s breakthrough raises urgent questions about the effectiveness of US export controls and the broader security implications of China’s AI progress. Can Washington and Europe really curb China’s technological rise through exports bans alone? Or do these policies rather risk accelerating China’s self-sufficiency while leaving the West vulnerable to AI-driven security threats?
Since 2022, the US has tightened restrictions on China’s access to cutting-edge semiconductors, banning the export of high-end NVIDIA AI chips (A100 and H100) and limiting the sale of advanced semiconductor manufacturing equipment, including extreme ultraviolet (EUV) lithography machines. In theory, these measures were designed to delay and obstruct China’s AI progress and preserve US dominance. In practice, they have proven inadequate as China has demonstrated remarkable resilience and adaptivity:
- Without access to top-tier GPUs, Chinese AI firms have focused on software innovations that maximize the performance of older, domestically available chips. DeepSeek-R1, for instance, was trained on a distributed computing network, compensating for hardware limitations through sophisticated algorithmic efficiency.
- China imported an estimated 80,000 A100 GPUs before the full effect of US restrictions kicked in. Additionally, Chinese firms have sourced restricted chips through intermediary nations such as Singapore and the UAE, a loophole enabling them to bypass direct US trade bans so far.
- Unlike in the US, where AI development is primarily driven by private sector competition, China’s AI sector benefits from direct state investment. The China National AI Institute and state-backed firms have received billions in subsidies, ensuring continuity. On January 20, the day DeepSeek-R1 was released to the public, founder Liang attended a closed-door symposium for business people and experts hosted by Chinese Premier Li Qiang, according to state news agency Xinhua, reinforcing the vital role of the state.
Beijings’ workarounds highlight a significant flaw in US export control policies: broad restrictions alone cannot curb China’s progress. Instead, Washington and Europe must adopt a more fluid approach that blends containment with proactive technological investment and coordinated international strategy.
China’s AI advancements are not just an economic challenge; they pose direct security risks to the US and its allies. The intersection of AI with surveillance, intelligence gathering, and autonomous military systems is reshaping global security in ways that demand urgent policy responses. Recent reports revealed that China has expanded its signals intelligence (SIGINT) operations in Cuba, using AI-driven analytics to enhance surveillance of US military activities. This suggests that AI is increasingly used to augment China’s global intelligence capabilities, making traditional counterintelligence measures less effective. Furthermore, the International Institute for Strategic Studies (IISS) predicts that China’s military AI will play a critical role in future combat operations, particularly in autonomous naval and air systems.
This growing gap has not gone unnoticed in the private sector. Dario Amodei, CEO of the US AI start-up Anthropic and Chair of the US-China AI Dialogue, has warned that China’s advancements in AI could quickly extend beyond civilian applications, warning that “It seems likely that China could direct more talent, capital and focus to military applications of the technology. Combined with its large industrial base and military-strategic advantages, this could help China take a commanding lead on the global stage, not just for AI but for everything” (Financial Times, 2025). With the US Air Force only beginning to integrate AI into its operational structures– expecting entire AI-assisted drone squadrons by 2050– the technological gap in military AI between China and the US may be wider than Washington assumes.
China’s AI resilience has exposed the weaknesses of a US strategy that relies too heavily on export controls. While restrictions remain an important tool, they must be complemented by proactive investment, stronger intelligence partnerships, and a competitive strategy that ensures technological leadership in AI security rather than merely trying to slow down China’s progress. It is simply not enough to bar China from modern technology; Europe and the US must coordinate and cooperate better to defend their technological advantage.
Deepseek’s “Sputnik” moment should serve as a wake-up call: AI is not just a commercial race but a security imperative. The US and Europe must act decisively, integrating their AI strategies to maintain control over artificial intelligence’s future. If policymakers fail to adapt, they risk not only economic decline but also a world where AI power is concentrated in the hands of an authoritarian regime.
The question is no longer whether China can bypass restrictions but whether if the US and Europe will have the foresight to lead in AI security, governance, and innovation.
