‘RAMageddon’ fallout

US' technology-denial strategy is driving China's pursuit of endogenous capabilities to ease memory-chip shortage

The severe global memory chip shortage, dubbed "RAMageddon", is more than a supply-chain disruption linked to artificial intelligence demand: it exposes the limits of weaponized interdependence as a coercive strategy.

While many analyses find no quick resolution in sight as the shortage is structural, reflecting a reallocation of the memory fabrication base toward AI infrastructure that spurs consumption of memory chips by data centers, the geopolitical implication of RAMageddon lies in what it reveals about the United States' technology-denial strategy toward China.

Since 2022, the US has pursued escalating semiconductor export controls premised on what US scholars Henry Farrell and Abraham Newman term "weaponized interdependence" — the exploitation of chokepoints in global networks to exercise coercive power.

In October 2022, the US Bureau of Industry and Security issued sweeping export controls targeting advanced chips and chipmaking equipment destined for China. In October 2023, these rules were tightened further, closing loopholes that had allowed third-country rerouting and expanding the Entity List to cover additional Chinese AI companies. The 2024 amendments extended controls to high-bandwidth memory (HBM) specifically, while pressuring the Netherlands and Japan to impose restrictions on equipment exports to China by ASML and Tokyo Electron.

Early this year, the US imposed additional licensing requirements on manufacturing equipment related to dynamic random-access-memory (DRAM), potentially constraining Chinese producers' expansion plans.

The apparent logic of weaponized interdependence is straightforward: because advanced semiconductor manufacturing depends on irreplaceable nodes — ASML's extreme ultraviolet (EUV) lithography system, US-designed electronic-design-automation (EDA) software and the memory oligopoly itself — restricting China's access is intended to constrain its AI development. Interdependence thus becomes an asymmetric lever for geopolitical control.

However, the phenomenon of RAMageddon challenges that logic in two fundamental ways.

First, chokepoint control is a double-edged sword. As major memory giants redirect capacity toward a more lucrative AI-related industry, the resulting scarcity of conventional DRAM can harm the entire global electronics ecosystem, including US companies and allied economies. This is the "boomerang effect" in international relations: sanctions generate unintended costs that erode the sanctioner's own position. In this context, nine US industry associations warned the administration that the imbalance could raise the price of electronic goods for US customers.

Second, and more critically, the crisis creates the very market conditions that enable China's memory industry to achieve viability at scale — precisely the outcome that the US policy sought to prevent. By forcing incumbents to prioritize the high end, export controls and AI demand have jointly left openings in the middle of the market, offering ChangXin Memory Technologies (CXMT) and other manufacturers a strategic opening.

CXMT is not an isolated case. Across the semiconductor value chain, Chinese companies have leveraged exclusion as an accelerant. Huawei's semiconductor arm, HiSilicon, has pivoted to designing AI chips manufactured by SMIC on its most advanced available process node — currently 7nm using multi-patterning deep ultraviolet (DUV) lithography — circumventing the blockade of EUV lithography. SMIC has expanded capacity in Beijing and Shanghai, and its reported record revenue in 2025 makes it the world's third-largest foundry. Yangtze Memory Technologies Corp, despite being placed on the US Entity List in late 2022, now accounts for 13 percent of the global NAND flash market and reportedly supplies domestic smartphone makers.

The pattern is revealing: US restrictions have triggered a burst of new Chinese investment and market adaptation, contributing to the development of indigenous capability.

As US economist Albert Hirschman argued in his foundational work on the political economy of trade, economic coercion works best when the target lacks alternatives. When coercion persists over time, however, it incentivizes the target to develop substitutes, and those substitutes, once built, permanently reduce the coercer's leverage. This is precisely what is occurring in the memory chip sector.

The structural market crisis, especially the HBM-driven supply crunch, has created a demand vacuum that may lower the barrier to entry for new producers including those from China. Industrial policy matters, but market opportunity is the enabling condition.

The growth of CXMT, SMIC and Huawei maps precisely onto Hirschman's argument of how coercion erodes its own foundations. The dominant power's own strategic choices, including export controls and AI preference, interact with market dynamics to produce conditions that inadvertently accelerate the challenger's rise.

This phenomenon can be described as "displacement through a market structural crisis". The more aggressively the US pursues AI supremacy in memory-intensive computing, the more it starves the conventional memory market, and the more space it opens for Chinese producers. The definition of "strategic technology" thus requires dynamic and dialectic thinking.

US controls focus on the leading edge, such as EUV, HBM and advanced logic. But when conventional DRAM becomes scarce because the leading edge has absorbed most, if not all, capacity, the frontier of competition shifts. Controlling the top of the stack matters less if the system cannot supply the middle.

The shift in competition is already visible. The history of semiconductor technology shows that yesterday's leading edge becomes accessible through alternative engineering pathways: SMIC's achievement of functional 7nm chips without EUV lithography is the most striking recent demonstration. The risk for the US is not merely that it loses market share in conventional memory, but that the cumulative learning and capital formation enabled by the mid-market eventually propel Chinese companies toward the high end as well.

RAMageddon is more than a supply-chain crisis; it is a natural experiment in the limits of technological hegemony, demonstrating that weaponizing chokepoints generates second-order effects that can undermine the original strategic intent.

For China, the path to semiconductor self-sufficiency need not run through the most advanced nodes first; it can run through the vast, underserved middle of the market. For the US, the lesson is revealing: a strategy predicated on permanent technological denial must contend with the possibility that denial itself creates the conditions for displacement. The question is no longer whether China has to be at the top in the global memory industry, but how quickly the displacement will happen. Continued rigidity in US export controls will not necessarily reverse China's progress, but will only ensure that the transition occurs on terms less favorable to US industry and consumers alike.

The author is an assistant professor at the Institute of International and Regional Studies at Zhejiang University of Technology, a researcher at Zhejiang Development and Security Research Think Tank Alliance, and a policy consultant for the OECD.

The author contributed this article to China Watch, a think tank powered by China Daily. The views do not necessarily reflect those of China Daily.

Contact the editor at editor@chinawatch.cn.