China is projected to exceed an 80% self-sufficiency rate in artificial intelligence (AI) chips within three years. While the United States’ stringent semiconductor export controls delivered a critical blow targeting China's AI Achilles’ heel, they have paradoxically become a catalyst for boosting the strength of China's semiconductor ecosystem. China is working to turn its blueprint for 'technological independence' into reality by establishing data centers powered by its own AI chips and unveiling humanoid robots equipped with domestically produced graphics processing units (GPUs).
Global investment bank Morgan Stanley predicted in its latest report titled 'China AI: The sleeping giant awakens' that China's self-sufficiency rate for AI chips, which was at 34% as of last year, will surge to 82% by 2027. Without U.S. sanctions, China would have continued to rely on overseas chips, such as those from NVIDIA. The firm analyzed that, under external pressure, China has launched an all-out effort toward complete independence and is building a self-sufficient ecosystem faster than expected.
The biggest driving force behind explosive growth is 'talent.' About half of the world’s AI researchers are based in China. Leveraging this, China has pursued a systematic strategy to build a self-sufficient ecosystem. As soon as the supply of top-tier chips was obstructed, China secured large quantities of downgraded versions of NVIDIA’s flagship AI chips, such as H20, while concentrating on enhancing software technologies to maximize performance by bundling multiple units. The open-source model DeepSeek-R1 has become a symbol of the Chinese solution of overcoming hardware disadvantages with software ingenuity, implementing AI at a cost far lower than GPT-4o.
China has also loosened funding. It has poured enormous amounts of research and development (R&D) funds into its AI ecosystem and has supported domestic corporations through 'priority purchasing by public institutions' by leveraging its vast domestic market. Huawei is challenging NVIDIA by showcasing an AI model training system equipped with over 10 trillion parameters through a 'supercluster' composed of thousands of its self-developed AI chips, 'Ascend 910.' China’s largest foundry company, Semiconductor Manufacturing International Corporation (SMIC), has surpassed the barriers of the 7-nanometer process and has converted Huawei's designed AI chips into reality. There is an evaluation that a virtuous cycle of self-sufficiency has been completed, leading from 'utilization of alternatives → creation of a domestic market → acceleration of self-development → mass production.'
The independence of the 'AI brain' is shaking the foundations of future industries beyond mere chip production. The humanoid robot market, projected to grow to $5 trillion (approximately 6,900 trillion won) by 2050, is a prime example. Morgan Stanley foresees that China will account for 30% of the global humanoid robot supply, leveraging the cost competitiveness gained from in-house sourcing of AI chips and key components. Estimates suggest that manufacturing costs for humanoid robots produced using the Chinese supply chain will be merely one-third of those within the global supply chain.
Companies leading China's AI surge are widely spread across various industries. Not only Huawei and SMIC but also Alibaba and Tencent, which are at the forefront of the AI platform competition, as well as robotics appliance company Ecovacs and autonomous driving solutions firm Horizon Robotics, are accelerating AI innovation, according to Morgan Stanley. Through this robust ecosystem, China's core AI industry is expected to grow to 1 trillion yuan (approximately 190 trillion won) by 2030.
The case of China demonstrates that the success equation in the AI power struggle is changing. Rather than the conventional belief that securing the highest-spec semiconductor chips directly correlates with market dominance, the 'power of the ecosystem'—which effectively intertwines somewhat scarce hardware with software and systems to create value—has become a key variable in competition.
Artificial intelligence was also a major topic in South Korea's 21st presidential election. However, it is regrettable that the focus was more on how much taxpayer money would be invested in the short term rather than a long-term vision for building a sustainable ecosystem. We can only hope that the newly elected government will work strategically to attract talent, support corporations, and foster a technological ecosystem that will serve as a future growth engine.