The battle over U.S. export controls on artificial intelligence (AI) chips has reached a fever pitch, with the semiconductor industry, policymakers, and global tech giants all pulling in different directions. As the annual defense policy bill winds its way through Congress, a controversial provision targeting AI chip exports to China has become a flashpoint, pitting national security concerns against economic interests and technological rivalry.
On September 7, 2025, the Semiconductor Industry Association (SIA), representing the nation’s major chipmakers, intensified its lobbying efforts against a measure championed by Senator Jim Banks. The SIA urged House and Senate leaders to strip Banks’ proposal from the defense bill, arguing that further tightening export controls would do more harm than good for American companies. According to the SIA, current restrictions already pose significant challenges, and additional constraints could jeopardize the competitiveness of U.S. firms in the global marketplace.
This lobbying push comes on the heels of a high-profile deal struck in August between the Trump campaign and leading chipmakers Nvidia and AMD. The agreement, as detailed by The Wire China on September 7, allows these companies to sell certain lower-grade computer chips to China, provided the U.S. government receives a 15 percent cut of the proceeds. The arrangement marks a notable shift from the U.S. government’s six-year trajectory of steadily tightening chip export restrictions, a policy that began during Donald Trump’s first term and continued under President Joe Biden.
The new deal has sparked fierce debate in Washington and beyond. Some policymakers and analysts see it as a dangerous concession that could undermine U.S. national security. Others argue it represents a pragmatic compromise, balancing the need to maintain leverage over China with the economic reality that American chipmakers are eager to retain access to the lucrative Chinese market. As The Wire China put it, the central question is whether the U.S. should keep China hooked on its chip technologies—thus preserving influence and revenue—or clamp down harder, risking a technological “Sputnik moment” that could spur China to leapfrog U.S. capabilities.
“Are the Nvidia and AMD deals really as detrimental to U.S. national security interests as many China hawks fear?” The Wire China asked in its cover story. The answer, it seems, depends on whom you ask. Proponents of stricter controls warn that even limited chip sales could help China advance its AI ambitions, potentially threatening U.S. interests in areas ranging from military technology to economic competitiveness. Opponents counter that shutting China out entirely could backfire, prompting Beijing to double down on developing its own alternatives—and costing U.S. firms billions in lost sales.
Recent developments in China suggest that the latter risk is far from hypothetical. In a September 7 opinion piece, The Wall Street Journal highlighted Beijing’s aggressive push to localize its AI chip supply chain. The Chinese government has directed local data centers to allocate at least 50 percent of their capacity to domestic chips, and the country’s leading open-source AI models are now optimized to run on local silicon. Even more striking, China’s foundry technology is reportedly advancing faster than many in Washington had predicted.
Take Huawei, for example. Earlier industry forecasts estimated the tech giant would produce no more than 200,000 chips this year. But according to updated figures cited by The Wall Street Journal, that number is now closer to one million—a fivefold increase. “Far from limiting Beijing, export restrictions have created conditions for faster localization and independence,” the article argued. Export-control advocates, the piece contended, often confuse the roles of fabless chip designers (like Nvidia and AMD) and foundries that manufacture chips. Regulating the end product, it suggested, isn’t the same as controlling the means of production.
For American chipmakers, the stakes are enormous. China represents not only the world’s largest AI talent market but also a massive customer base for semiconductors, manufacturing equipment, and design software. “Export controls on GPUs punish American firms and create a stronger incentive for China to develop an alternative ecosystem,” The Wall Street Journal warned. At the same time, these restrictions deny U.S. companies access to critical growth opportunities in China, potentially undermining their ability to invest in future innovation.
All the while, the broader geopolitical context is shifting. As The Wire China reported, the U.S. has spent years pressuring allies—most notably the Netherlands and Japan—to adopt similar restrictions on chip sales to China. Yet, as foreign-invested companies continue to play a major role in both countries’ export economies, the effectiveness and sustainability of this strategy remain open questions. Meanwhile, the September 1 meeting between Chinese, Russian, and Indian leaders in Tianjin offered little reassurance that the world’s major powers are finding common ground on technology or trade.
The debate over chip export controls is also sparking reflection on the fundamental differences between American and Chinese approaches to technology and governance. In his new book, Breakneck, former technology analyst Dan Wang contrasts America’s “legalistic procedure fetish” with China’s obsession with outcomes. Wang suggests that China’s ability to rapidly build infrastructure—and, by extension, its tech sector—may give it an edge in the race for AI supremacy. But he also notes the unpredictability of policymaking under the Chinese Communist Party, where “the line between rationality and madness can be a pretty thin one.”
Back in Washington, the path forward remains uncertain. The defense policy bill—and the fate of Senator Banks’ export control provision—hangs in the balance. The SIA and its allies are pulling out all the stops to persuade lawmakers that a more flexible approach is needed. At the same time, national security hawks are warning that any retreat could prove disastrous in the long run.
As the debate rages, one thing is clear: the intersection of technology, trade, and national security has never been more contentious. The decisions made in the coming weeks could shape not only the future of the semiconductor industry but also the broader contours of U.S.-China relations—and, by extension, the global balance of power in the age of artificial intelligence.
The world is watching as Congress weighs its options, with industry leaders, policymakers, and international partners all anxiously awaiting the outcome. Whether the U.S. chooses to double down on restrictions or seek a new equilibrium, the ripple effects will be felt far beyond Washington and Beijing.
