The relationship between the U.S. and China in the technology sector has resembled an obstacle course for years, but in recent months, the dynamics appear to have reached a new peak. The latest export restrictions on semiconductors and advanced Nvidia chips to China, imposed by U.S. President Donald Trump, seem at first glance like another attempt by Washington to curb China’s rise in AI. However, considering how things have unfolded over the past few years, one must ask whether this strategy is already too late. China appears to have responded with its own progress, unveiling ambitious projects and AI models that could leave the U.S. without its long-held technological edge.
The primary target of American restrictions is currently Nvidia, a company valued at nearly $3 trillion and one of the most important GPU manufacturers in the world. These chips’ ability to “train” massive AI systems has made them a critical component for companies and institutions in China. However, the Trump administration introduced a licensing requirement for the sale of one of Nvidia’s latest chips, the H20, to limit or at least delay its delivery into Chinese hands. As a result, Nvidia’s stock took a noticeable hit, and the chip industry once again found itself in a state of uncertainty.
This technological context dates back to Trump’s first term, when shipments of advanced chips to Chinese tech giant Huawei were halted. At the time, Washington aimed to restrict access to key components for 5G networks, hoping to preempt China’s development of new technologies. The Biden administration expanded these measures, and now, with Trump’s return to the White House, it appears the bans will continue to accumulate. The problem for the U.S. is that China has used every opportunity in the meantime to strengthen its own capabilities, proving that even multiple restrictions cannot completely halt its progress.
One key example of China’s resilience and innovation is DeepSeek, an AI model recently introduced that has caused a considerable headache for U.S. strategists. The R1 version of this model has proven to be highly advanced, and claims that it was trained using a large number of legally purchased Nvidia chips only further confirm that sanctions have not fully achieved their intended effect. At some point in this ongoing cat-and-mouse game, the U.S. tried to fine-tune each new restriction to include specialized chip versions. However, engineers in China were able to exploit allowed “loopholes” in the rules, obtaining various components and assembled chips before the bans officially took effect.
This situation has naturally prompted Chinese companies to double down on developing their own domestic chips. Huawei, in particular, has stood out: rumors suggest the upcoming launch of their new Ascend 910c, followed closely by the more advanced 910d, which is expected to seriously challenge American alternatives. China is trying to reduce its heavy reliance on foreign technology, although it still uses elements that are difficult to simply “disconnect” from the global supply network. Nevertheless, the massive financial and human capital that China is investing daily into AI and semiconductors is a clear warning to the West that a return to the previous balance may no longer be possible.
The fact remains that the U.S. still possesses vast reserves of the most advanced technologies, but neither Trump nor his predecessors have succeeded in halting the rise of global competition. In fact, attempts to isolate China from Western chips have, in some ways, further motivated it to accelerate its own R&D. One could even say that necessity has pushed China to become more innovative, particularly in finding alternative ways to build AI models. In this context, DeepSeek has become not just a technical achievement, but also a symbolic one, signaling that Chinese firms can progress despite international obstacles.
Some companies and experts in the U.S. believe that these new restrictions may be helpful only in the short term, slowing Chinese AI models just enough to create a public perception of control within America. However, the tech war is not fought in quick moves—it stretches over the long term, and so far, China has demonstrated impressive persistence and flexibility. Its independent progress in semiconductor and algorithm development could eventually surpass any Western advantage gained through current U.S. dominance in the chip market.
It’s worth noting that China’s AI focus is not only economic, but also military—especially through its policy of “military-civil fusion.” In theory, the most advanced commercial technologies can be leveraged for the People’s Liberation Army. This is the root of Washington’s fears: all AI advancements—from logistics to drone applications—can be used to enhance defense or project power. For American strategists, restrictions are therefore seen as necessary—even if they only buy a little time, that small margin could be decisive.
Critics of U.S. policy, on the other hand, emphasize that the key issue is long-term security and economic sustainability. Not only could companies like Nvidia lose a large portion of the Chinese market, but this might also reduce their research and development budgets. Ultimately, such a blow could undermine American innovators just when they need to make big investments to stay ahead of Beijing. Moreover, the growing wall of mistrust between the U.S. and China casts doubt on any serious cooperation in AI safety—potentially dangerous in a world where AI is rapidly advancing and its power extends beyond any single continent.
For now, Trump’s proposals show no signs of easing toward China, and continued American lobbying for allies in Europe and Asia to adopt similar restrictions points to further escalation. However, this approach is at the very least challenging: countries placed in the “second” or “third” tier for access to U.S. chips may start questioning whether strict adherence to restrictions is in their best interest. In upcoming negotiations, the Trump administration will likely try to fine-tune the rules again, but whether that will truly prevent China from becoming a technological powerhouse remains highly uncertain.
In the end, there is little doubt that the U.S. and China are entering a phase of intensified competition in AI capabilities. China’s vast market, momentum in development, and growing support from both the state and private sector could speed up its rise to the top—perhaps more quickly than anyone in Washington expects. The American strategy is based on extending its current lead, but in a world of rapid innovation, that’s hard to sustain without active engagement and consistent, not just reactive, measures. On the other hand, if China indeed has everything it needs—and recent developments suggest it does—then even reinforced restrictions may only delay the inevitable.
About the Author
