The rivalry between the United States and China over artificial intelligence (AI) and semiconductors is often described as a race. In reality, it resembles a strategic dance—one step forward, one step back, and constant recalibration. For years, Washington moved aggressively to restrict Beijing’s access to advanced chips, hoping to slow China’s progress in AI and high-performance computing. Yet these restrictions had unintended effects: U.S. businesses suffered revenue losses, China doubled down on self-reliance, and alternative supply chains emerged.
Now, the U.S. is quietly changing course. The Trump administration recently approved Nvidia’s H20 chip exports to China under a revenue-sharing system, requiring the company to hand over 15% of proceeds to the U.S. government. Washington also signaled flexibility by allowing downgraded versions of Nvidia’s Blackwell chips to reach Chinese buyers.
Read More: https://www.moneynewsweb.com/india-set-to-overtake-as-second-biggest/
This adjustment reflects an important recognition: blanket bans may have backfired. Instead, policymakers hope selective dependence will keep China tethered to American technology, maintaining leverage in the long run. But the critical question remains—has this shift come too late?
Washington’s Policy Reversal
The U.S. semiconductor strategy has shifted in tone and execution over the past three years. Between 2022 and 2023, sweeping restrictions blocked China from acquiring advanced processors such as Nvidia’s A100 and H100, chips essential for training large AI models. The aim was clear—deny China the tools it needed to dominate AI. But the results were more complex.
Economic blowback: China represents about one-third of global semiconductor demand. Cutting off this market disrupted American chipmakers’ business models. Nvidia’s CEO Jensen Huang revealed that the company’s market share in China plunged from 95% to 50% after the Biden administration’s restrictions. He estimated the H20 export ban alone cost Nvidia nearly $15 billion in lost revenue.
Fueling Chinese innovation: Instead of halting progress, sanctions accelerated Beijing’s pursuit of independence. Supported by massive state subsidies, Chinese firms like Huawei and SMIC raced to improve chip design and fabrication. Huawei’s Ascend 910C processors and SMIC’s progress toward 5-nanometer manufacturing demonstrate how far China has come, even if it still trails industry leaders like TSMC. By some estimates, domestic AI chips could capture more than half of China’s market share by 2027.
Recognizing these realities, the Trump administration pivoted. It approved limited exports of Nvidia’s H20 and AMD’s MI308, hoping to prevent China from fully severing ties while preserving U.S. companies’ foothold in the global market. Commerce Secretary Howard Lutnick summed up the strategy: “We want to keep the Chinese using the American technology stack, because they still rely upon it.”
However, this reversal has sparked fierce criticism. Lawmakers on Capitol Hill argue that loosening restrictions jeopardizes national security. Proposals such as the Chip Security Act (S. 1705 | H.R. 3447) would mandate tracking systems inside exported chips, allowing Washington to monitor their use in China. While technically challenging and commercially costly, the idea underscores how mistrust shapes U.S. policymaking. When Congress reconvenes in September, calls for tighter controls will likely intensify.
Beijing’s Dual Response
China’s reaction to Washington’s recalibration has been both pragmatic and strategic.
Long-term independence: Chip self-reliance is a central plank of China’s national security strategy. Initiatives like Made in China 2025 and successive five-year plans channel billions toward domestic innovation. Huawei’s Ascend 910C chip, scheduled for mass production in September 2025, rivals Nvidia’s H20 in AI inference tasks and even outperforms it in energy efficiency. Though built using SMIC’s 7nm process—three years behind TSMC’s cutting edge—the chip’s capabilities highlight how rapidly China is catching up. Combined with Huawei’s CloudMatrix 384 system, which relies on optical networking, these advances demonstrate China’s steady march toward technological independence.
Short-term reliance: Despite progress, Chinese firms continue to buy U.S. chips in bulk. After restrictions on the H20 eased in mid-2025, Chinese tech giants reportedly ordered at least 700,000 units. The reasons are clear: Nvidia’s chips remain superior in training efficiency and are deeply integrated into the CUDA software ecosystem. Competing Chinese chips often face ecosystem gaps and higher maintenance costs—up to 50% more in some cases.
Rising caution: At the same time, Beijing is wary of hidden risks. In July 2025, China’s Cyberspace Administration (CAC) reportedly questioned Nvidia about potential backdoors in the H20. Officials also summoned companies like Tencent and ByteDance, urging them to justify purchases and prioritize local alternatives. The suspicion that U.S. chips could include surveillance or tracking systems reinforces the view that dependence on American technology is a strategic vulnerability.
This dual strategy—purchasing U.S. chips when necessary while accelerating independence—captures the paradox of China’s position. The more Washington tries to manage dependence, the stronger Beijing’s determination to eliminate it becomes.
Too Late for a Rethink?
The U.S. hopes its new policy will preserve leverage by ensuring China’s continued reliance on American technology. Yet momentum points in a different direction. The world is heading toward two parallel technology ecosystems—one led by the United States and its allies, the other by China and its partners.
China may not yet lead in absolute cutting-edge performance, but its progress is undeniable. Domestic chips are increasingly good enough for most applications within China and across parts of the developing world. Self-reliance is no longer just an economic ambition; it is now a national security imperative. Washington’s decision to attach conditions to exports, including potential tracking mechanisms, only deepens Beijing’s distrust.
Meanwhile, fragmentation is reshaping the global landscape. U.S. allies such as Japan, South Korea, and much of Europe will likely remain aligned with Washington’s standards. But in the Global South, Chinese chips offer affordable alternatives, even if they lag slightly in performance. Countries like Malaysia may straddle both ecosystems, absorbing the costs of dual compliance to maintain flexibility.
For global businesses, this divide means higher costs and greater complexity. Companies will face duplicated supply chains, diverging compliance rules, and software incompatibilities. Cybersecurity risks will multiply, while the efficiencies of a unified global tech order continue to erode.
Washington’s recalibration might delay China’s rise but cannot reverse it. The real challenge is not preserving total dominance but adapting to a world where technological supremacy is shared rather than singular. As Brookings scholar Ryan Hass has noted, the issue is less about securing a decisive lead and more about managing coexistence.
Frequently Asked Questions:
Why are semiconductors so important in the US-China rivalry?
Semiconductors power AI, smartphones, cloud computing, defense systems, and nearly every modern technology. Control over chip design and manufacturing directly impacts economic competitiveness and national security, making them central to the US-China rivalry.
What was the goal of US restrictions on Chinese chip access?
Washington aimed to block China from acquiring advanced processors like Nvidia’s A100 and H100, which are vital for AI development and supercomputing. The strategy sought to slow Beijing’s technological progress and protect US dominance in critical industries.
Did US sanctions on China’s chip industry work?
The sanctions had mixed results. While they limited China’s immediate access to advanced chips, they also hurt US companies like Nvidia and accelerated China’s efforts to build its own semiconductor ecosystem.
Why is Washington now easing export rules for some chips?
The US realized that blanket bans were backfiring—pushing China toward independence while harming American companies. By allowing limited exports under strict conditions, Washington hopes to keep China dependent on US technology.
How has China responded to US chip restrictions?
China has adopted a dual strategy. In the short term, it continues buying US chips to maintain competitiveness. In the long term, it is investing heavily in self-reliance, with companies like Huawei and SMIC developing advanced chips to reduce dependency.
What risks do US companies face in this rivalry?
Firms like Nvidia and AMD face revenue losses, shrinking market share in China, supply chain disruptions, and the risk of Chinese competitors eventually replacing them in global markets.
Could the world end up with two separate chip ecosystems?
Yes. Experts warn of a fragmented future where one ecosystem is led by the US and its allies, and the other by China and its partners. This split could increase costs, reduce efficiency, and create compatibility issues worldwide.
Conclusion
The U.S.-China battle over chips and artificial intelligence is no longer a simple contest of winners and losers. Washington’s attempt to block China’s access to advanced semiconductors slowed progress in the short term but also triggered unintended consequences—hurting U.S. companies, driving Chinese self-reliance, and accelerating the fragmentation of global supply chains. The recent policy shift toward selective engagement reflects a recognition that total bans were unsustainable. Yet timing matters. China has already made significant strides in chip design and AI innovation, supported by massive state investment and national security imperatives. As a result, the global technology order is heading toward two parallel ecosystems, each shaping markets, standards, and alliances.