The recent H200 chip export authorisation marks a new form of power signalling in the technopolitics sphere. Advanced AI chips are the newest weapons in the US-China great power rivalry, playing out in the battlefield of technological advancement and control. In the case of AI chips, state behavior shifts from the purely economic considerations typical of economic statecraft. Researchers have classified this as a separate category known as techno-statecraft or techno-economic statecraft. American strategy lies in supply chain diversification, machinery export control, and geopolitical lobbying to hit Chinese semiconductor financing. Its most potent strategic maneuvre lies in cutting off EUV lithography technology access, crucial to the production of AI chips – essentially China’s Achilles heel. Chinese strategy lies in production self-sufficiency through domestic impetus. Essentially, techno-statecraft is an emergent pattern in US-China state behavior, underscoring the political power of the AI chip – and the lack of it.
Sub 7 nm chips, i.e. semiconductor chips thinner than 7 nanometers, are significant components in AI manufacturing, used for computing and processing efficiently. NVIDIA’s H200 chip falls within this category. The process of manufacturing the chips themselves requires extreme ultraviolet (EUV) lithography technology, the sole producer of which is ASML Netherlands. The U.S. government lobbied for restrictions to the sale of EUV technology to China in 2024, effectively cutting off China’s sub 7 nm chip production capacities. This move forced Chinese AI companies such as Alibaba and ByteDance to fulfill their AI chip requirements from third party manufacturers. These circumstances form the aforementioned ‘Achilles Heel’ – the EUV lithography restrictions force Chinese AI industries, valued at over USD 175 billion in 2025, to move beyond domestic chip producers for their requirements. This maneuvre dually cripples Chinese AI chip manufacturers by taking away their primary market and exposes AI companies to vulnerabilities by increasing their reliance on external suppliers.
American strategy, while pursuing an aggressive stance in cutting off Chinese AI chip production at its roots, acknowledges US-China semiconductor chip interdependency in other technological milieus. In order to combat this, American markets have been derisking from China by diversifying to Vietnamese chip supply and removing Washington’s export controls to Vietnamese entities. These export controls have in turn been strengthened to further squeeze Chinese access to crucial AI chips, such as the NVIDIA H200 chip. This is purportedly done to avoid the dual use of the chip – for military uses rather than civilian – to threaten US security. The current H200 authorisation comes with a history of caveats, making American leverage over Chinese AI industries palpable. In addition to this strategy, the US also restricted the shipping of its machinery, which companies such as TSMC and SK Hynix use for chip production, to Chinese fabrication facilities over the past year. This multi-pronged techno-statecraft approach, consisting of cutting EUV lithography access, supply chain diversification, and stringent export controls, underscores the sheer power that the AI chip holds as one of the newest weapons in the US-China tech race.
In order to counter the effects of these restrictions, Beijing implemented a series of domestic policies to reduce dependence on foreign suppliers and increase self-sufficiency. This is demonstrated by the lack of H200 chip orders from Chinese AI companies. Another measure of self-reliance in the global supply chain is the 50% domestic manufacturing policy, requiring companies to use at least 50% domestically manufactured equipment. This policy, coupled with the “Whole-of-nation innovation” program to pool domestic human resources in strengthening semiconductor production capacities. Specifically in relation to advanced AI chips, China has been researching upgrading available Deep Ultraviolet (DUV) lithography technology from ASML to boost sub 7 nm chip production, with some successes. This strategy helps in derisking manufacturers such as TSMC and combatting export control policies from the U.S. and Taiwan. However, given that EUV technology is the industry standard and DUV machinery is not as time or quality-efficient, this still leaves the Chinese AI industry in a state of dependency.
The behavior of both parties speaks to underlying economic considerations, but it also strays into the political psychology of technological advancement and control. American diversification strategy deviates from the immensity of the Chinese market. Chinese narrative strategy on chip self-sufficiency seeks to nationalise chip production rather than diversifying with its southeast and east Asian counterparts, choosing to stay away from placing its eggs in different baskets. Despite domestic requirements for the H200 chip, Chinese companies have not been placing orders to avoid external dependency. These developments, while entangled with considerations of leverage and dependency, are inherently necessitated by techno-statecraft – promoting the advancement of technology and spearheading its control. The American-backed EUV lithography bottleneck to Chinese industries remains an efficient tool of techno-statecraft, uncovering a multi-faceted network of interdependencies and highlighting the significance of the AI chip in demonstrating and maintaining power imbalances between the US and China.
