The fight for dominance in artificial intelligence (AI) cannot be divorced from the global competition for cutting-edge semiconductors, the critical hardware behind advanced AI models. The United States and China, the world's two largest economies, have been locked in a high-stakes "chip war" for decades with each trying to achieve technological superiority. During President Donald Trump's second term, initiated in January 2025, the war has developed new turns, marked by a combination of pragmatic deal-making, enhanced controls, and realignment of global alliances. The article discusses the ways in which the US-China chip war is proceeding under Trump, highlighting the recent developments, their meaning for AI innovation, and the broader geopolitical implications.
The Stakes: Why Chips Matter for AI
The focal point of the US-China race in technology is the semiconductor sector, precisely the advanced chips—such as Nvidia's A100, H100, and Blackwell GPUs—employed for training and deploying such deep AI models like those powering ChatGPT, DeepSeek's R1, and other cutting-edge systems. These chips provide us with raw computing muscle needed to crunch huge volumes of data and hone advanced algorithms, and are a strategic asset of the type of oil in the 20th century. Whoever controls their production, diffusion, and innovation dictates not only economic prosperity but also military and geopolitical influence, as AI systems become ever more prevalent in surveillance, cyber warfare, and autonomous military systems.
The US has historically had the benefit of technology due to companies like Nvidia, AMD, and Intel, as well as its leadership in chip design and manufacturing gear. Taiwan's TSMC, a key US ally, produces a majority of the world's advanced chips, with Dutch company ASML providing critical lithography equipment. China, however, has been quickly bridging the gap with state-led efforts like its 2017 Next Generation AI Development Plan to make China global AI leader by 2030. By US export controls or otherwise, Chinese firms like Huawei and SMIC have been aggressively working on developing local alternatives, challenging the long-term sustainability of restrictions.
Trump's Approach: A Transactional Shift
The Trump government has taken a different approach in the chip war, coupling aggressive China restrictions with an otherwise negotiable deal-based approach to allies and neutral states. This is a divergence from the policies of the Biden government, which relied to a significant degree on across-the-board export controls, including the January 2025 AI Diffusion Rule, which was roundly criticized. That regulation divided the world into three groups, with virtually unlimited chip access for close friends like Canada and Germany, strict licensing to over 150 countries, and banning sales to adversaries like China. Industry giants like Nvidia and Oracle were joined by friends like Israel and Poland in condemning the rule as stifling innovation and pushing neutral countries towards Chinese alternatives. Trump canceled the AI Diffusion Rule in May 2025 for being too complicated and bureaucratic.
The Commerce Department outlined plans to replace it with a country-by-country global licensing scheme, aiming to use chip access as a trade and diplomatic club. It is part of Trump's transactional foreign policy, which values bilateral agreements over multilateral frameworks. As an example, during a May 2025 tour of the Middle East, Trump reached agreements to send hundreds of thousands of Nvidia and AMD chips to Saudi Arabia and more than a million to the UAE, primarily for US business ventures like OpenAI. The deals would increase US leverage in the region and counter China's growing economic presence in the region. But this approach has caused tensions within the administration.
China hardliners, concerned with Gulf states' relations with Beijing, worry that advanced chips would be sent to China indirectly. The Commerce Department countered by sending warnings against "diversion tactics" and tightening limits on using US chips to train Chinese AI models. Additionally, the administration declared that using Huawei’s Ascend AI chips anywhere globally could violate US export controls, signaling a hardline stance on Chinese technology. China’s Response: Innovation and Workarounds
Despite US efforts to curb its access to advanced chips, China has shown resilience. Chinese tech giants like Alibaba, Xiaomi, and DeepSeek have invested heavily in AI, with DeepSeek’s R1 model—rivaling OpenAI’s o1 at a lower cost—demonstrating that China is closing the gap. DeepSeek reportedly leveraged a stockpile of 10,000 Nvidia A100 GPUs acquired before 2022 export controls, highlighting how Chinese firms have exploited loopholes, such as leasing offshore servers or purchasing chips through intermediaries.
Huawei, which has been targeted for US sanctions since 2019, is about to reveal its Ascend 910c and 910d AI chips that are able to match Nvidia's H100. The actions reflect not only sanctions avoidance but also building self-sufficiency. One of the leading AI specialists, Kai-Fu Lee, indicated Chinese AI firms now trail their American counterparts by only three months in core technologies, significantly better than before, driven by urgency and government support.
China has also responded with strategic measures, including restricting access to critical minerals used to make chips that could drive up costs for global manufacturers such as TSMC. The tit-for-tat action mirrors the broader trade tensions, though a May 2025 90-day tariff ceasefire suggests a short-term defusing.
Implications for the Global AI Race
Trump's actions are remapping the global AI map.
Through the repeal of the AI Diffusion Rule and pursuing bilateral agreements, the US is betting on its tech superiority for economic and diplomatic gain. This might, however, push allies away who prefer stable, multilateral deals and are not keen on taking risks with the US, and might see neutral nations turn towards China if access by the US to chips has too much baggage. The EU, for instance, is exploring a chips alliance to reduce dependence on US technology, with Southeast Asian nations walking the US-China fence carefully. The removal of the AI Diffusion Rule has alleviated export burdens for US companies, with firms such as Nvidia freely exporting chips to countries such as India and Saudi Arabia without additional licensing.
But tighter controls on China, such as the ban on Nvidia's H20 chip, have cost it billions, with corporate profit versus national security trade-offs coming into stark relief. Nvidia CEO Jensen Huang has warned that too tight controls threaten losing overseas markets to Chinese competitors, a concern shared by industry leaders. China’s progress, meanwhile, suggests that export controls alone cannot maintain the US lead. DeepSeek’s success and Huawei’s chip advancements indicate that innovation, spurred by restrictions, is narrowing the gap. Experts argue that the US must invest heavily in R&D, STEM education, and domestic AI infrastructure—such as the proposed $500 billion Stargate project with OpenAI, SoftBank, and Oracle—to stay ahead.
The Road Ahead
The Trump China chip war is a careful balancing act.
While on the one side, the administration is doubling down on denying China access to cutting-edge technology, targeting firms such as Huawei and constricting supply chain control, it is simultaneously opening doors for allies and swing states to access US chips and counter China's influence while energizing American industry. This two-pronged approach is dangerous: loosening restraints for the Gulf states may lead to technology leakage, while overly ambitious restrictions may trigger China's autarky and isolate global allies. The larger question is whether the US can maintain its AI edge through the imposition of export controls or whether it must out-innovate China. As Peterson Institute's Martin Chorzempa noted, "You can't really gatekeep" in an open-source world where knowledge is freely shared. Trump's transactionalism may win short-term battles, but the long-term victory will depend on robust investment in domestic capability as well as strategic cooperation with allies.
In this evolving chip war, the stakes escalate above technology to geopolitics. The nation controlling the chips that power AI will determine the shape of economic and military hegemony. Currently, the US enjoys the advantage, but China's speeding strides and Trump's desperate negotiations are rendering the outcome far from certain.
Comments
Post a Comment