The key geopolitical battle of the twenty-first century is no longer about oil reserves, nuclear weapons, and battleships. The key battle is now being waged using artificial intelligence, semiconductors, and digital technologies. What the world is witnessing today is an AI Cold War between the United States of America and China, which is not only changing the nature of global power and global commerce but also globalization itself.
Artificial intelligence has grown into something more than a technological tool for businesses; today, it is viewed as a strategic resource. Governments around the world now see AI as central to military superiority, economic competitiveness, cyber capabilities, and geopolitical influence. The country that dominates AI will likely dominate the next era of global leadership. This realization has transformed the US-China AI race into one of the most consequential rivalries in modern history.
At the heart of this confrontation lies semiconductors, the advanced chips required to train and operate powerful AI systems. These semiconductors are further used to make a Graphics Processing Unit (GPU) and without high-performance GPUs, modern AI models simply cannot function. This is why Nvidia, once known mainly for gaming graphics cards, has become one of the world’s most strategically important companies.
Nvidia’s GPU chips power everything from OpenAI’s large language models to military-grade computing systems and this makes it the first ever company to reach USD 5 trillion market capitalization in 2025. Its dominance in the GPU market is so big that it controlled around 92 percent of the GPU market before 2026. The dominance of Nvidia in the GPU and the AI market has given Washington enormous leverage in the semiconductor war with China. Before export restrictions intensified, Nvidia reportedly controlled nearly 95 per cent of China’s advanced AI chip market. However, by 2025, its market share in China had fallen dramatically to less than 60 percent, while Chinese firms captured nearly 41 per cent of the domestic AI accelerator market.
The reason for this decline is not market competition alone, they are geopolitical also . Over the past several years, Washington has imposed sweeping chip export controls designed to restrict China’s access to advanced semiconductors and AI infrastructure. In October 2022, the United States announced two rules aimed at restricting China’s ability to obtain advanced computing chips, develop and maintain supercomputers, and manufacture advanced semiconductors . These controls were further expanded in October 2023 and December 2024. The Trump administration then went on to place further restrictions on the export of chips in March 2025 by placing numerous Chinese entities on a blacklist prohibiting their participation in trade in semiconductors and other sophisticated technologies. In fact, the US Department of Commerce candidly declared that these actions were taken as a means to ensure that China is not able to produce sophisticated semiconductors that would help it enhance its military and surveillance capabilities. Washington also pushed countries like the Netherlands and Japan to restrict export of semiconductors.
This strategy reflects a broader shift in global politics, that is technology restrictions are no longer isolated trade policies rather they have become instruments of economic statecraft.
Yet the strategy has produced mixed results. While export controls have undeniably complicated China’s access to advanced AI hardware, they have also accelerated Beijing’s push for technological self-sufficiency. At least nine Chinese companies, including Huawei’s Ascend, Baidu’s Kunlunxin and Alibaba’s T-head have significantly expanded domestic AI chip production and surpassed a significant threshold of 10,000 units in terms of order shipments.. According to the latest report, Chinese semiconductor companies shipped around 1.65 million AI accelerators in 2025, reducing the market leadership of Nvidia and boosting the homegrown ecosystem in China
The Chinese company Huawei stands out as the focal point of China’s plan to resist AI. The Chinese company is promoting its AI processors as replacements for restricted Nvidia products. According to the latest estimates, Huawei plans to produce around 600,000 Ascend 910C processors in 2026 as China accelerates efforts to reduce dependence on American technology.
Ironically, even Nvidia CEO Jensen Huang has warned that overly aggressive restrictions may backfire strategically. In recent interviews, Huang argued that pushing China out of the American semiconductor ecosystem could ultimately strengthen China’s domestic industry and weaken long-term US influence over global AI markets.
This also signifies one thing that the scale of this competition between the U.S and China is staggering. According to research on global AI supercomputers the computational performance has been doubling roughly every nine months, while hardware acquisition costs and power requirements have doubled annually. By 2025, the world’s leading AI supercomputer reportedly used around 200,000 AI chips and required energy equivalent to powering approximately 250,000 households.
In many ways, the AI Cold War is redefining the meaning of power in the twenty-first century. Unlike the ideological Cold War of the past, this rivalry is being fought through algorithms, semiconductors, data, and digital infrastructure. Though the semiconductor battle and the growing number of restrictions on chip exports might give China some temporary reprieve, they would simultaneously be contributing to a worldwide trend toward technological independence. The implications of a contest between the US and China to dominate in AI will go beyond technology into military calculations, economic relationships, and even the very survival of globalization itself.
