When the Chips Are Down

Given the complexity of the topic at hand, the CNAS Gaming Lab developed a strategy game to examine global semiconductor competition. Games provide a “safe to fail” environment, which is particularly conducive to examining poorly understood problems. Games also serve as powerful tools for establishing a shared understanding of a problem, given their collaborative and experiential format and ability to convene different communities. The Chips Are Down game enabled the CNAS team to learn more about the competition for semiconductors, while providing game participants with a shared understanding of the critical implications of the competition.

The Chips Are Down game produced critical insights into the nature of U.S.-China strategic competition and global competition for semiconductors, discussed in this report. This report first provides an overview of the game including its purpose, the scenario, and the game design. Next, it details four key insights derived from the game, examining their emergence during gameplay and their real-world implications. Lastly, it concludes with recommendations for overcoming a set of challenges stemming from these insights, aimed at improving the U.S. position in future strategic competition.

The Chips Are Down: The Game

CNAS conducted a virtual strategy game in April 2021 to better understand how China could attempt to influence and exert control over Taiwan’s semiconductor industry while remaining under the threshold of war and the subsequent implications for strategic competition with the United States. The game featured 30 high-level participants drawn from government, industry, and civil society. The purpose of the game was twofold: to understand how China could shape the semiconductor industry and strategic technology competition in its favor, and to investigate how the United States and Taiwan could counter such attempts. The goal was to identify areas of risk and vulnerability within the semiconductor industry and better understand the tradeoffs of different strategic approaches that the teams could adopt. More broadly, the aim was to glean insights about the nature of the U.S.-China strategic competition, with respect to Taiwan’s critical role.

The semi-structured game was conducted virtually and featured three teams: the United States (Blue team), China (Red team), and Taiwan (Green team). The U.S. team represented an interagency task force, while the China and Taiwan teams represented coordinating committees. The CNAS control team presented all three teams with a baseline scenario, but with different objectives that reflected the teams’ national interests (see Appendix A).

The game scenario began in January 2025, following a period of intensified strategic competition between the United States and China, focused on the economic domain and the competition for technology resources, ideas, and talent. Semiconductors, particularly leading- edge manufacturing capabilities, were a key focus, as the United States and China both sought to enhance their access to semiconductors by strengthening domestic design and fabrication capabilities. China doubled down on an ambitious technology indigenization strategy, and the United States offered incentives to expand fabrication capacity at home. Despite their efforts, both countries remained heavily dependent on Taiwan for leading-edge chips, and Taiwan became the flashpoint in the broader tensions between Washington and Beijing.

The competition over Taiwan’s semiconductor sector—as well as the broader U.S.-China competition for influence—came to a head after a political crisis between the two countries emerged over U.S. support for Taiwan. Shortly thereafter, three TSMC manufacturing facilities reported an issue in their manufacturing lines and halted all production. TSMC engineers discovered that code used to manufacture leading-edge chip designs was corrupted, although it was unclear whether it was due to software failure or a cyber attack. The result was a two month suspension in chip fabrication, creating a global shortage in leading-edge chips.

The teams were tasked with creating a strategy to respond to this crisis and gain the upper hand in the competition. To implement their strategy over the course of a one-year turn, the teams developed a set of diplomatic, informational, military, economic (DIME), and civil actions (see Appendix B). These actions were linked to targets, which could be people, organizations, or processes located in a particular country. The teams also were tasked with explaining the intended effect of their actions. For example, the China team could choose to conduct an offensive cyber operation (action) against TSMC foundry air filtration systems at its Hsinchu Science Park manufacturing facility (target) with the aim of thwarting fabrication of 2 nm chips (effect).

These actions sought to influence five key indices represented in the game. These included:

• Public sentiment, which represented public opinion and general support to the government and leadership
• Technology levels, which represented the amount of technology being used by a particular country
• Health, which tracks the financial health of companies located in a country
• Output, which represents the number of chips manufactured in a given year
• Demand, which represents the number of chips that customers in all sectors need and want.

Together, these indices represent a simplified qualitative model of the global semiconductor industry developed by the CNAS team. The components were semiconductor equipment manufacturers with a particular emphasis on the Dutch firm ASML; raw materials necessary for semiconductor fabrication with Japanese photoresists weighted more heavily; semiconductor design capability; memory chip production; a basic geographic breakdown featuring China, the United States, Taiwan, and rest-of-world; global semiconductor fabrication output; and the global customer base. The technology levels, the output to demand, and the health of companies represent how the model tracks the global semiconductor industry, while public sentiment tracked country stability and other domestic factors. The indices changed throughout the game as actions could improve or negatively impact them. The impact of the actions on select indices, such as public sentiment, had implications for a team’s resources and capacity. The success or failure of these actions was determined by the CNAS team, leveraging the set of rules it developed while creating a model of the semiconductor competition.

The teams could take actions independently or they could choose to cooperate with another team. Cooperative actions held a greater chance of creating an effect, given the shared resources and level of effort. Moreover, as certain actions—such as sanctions or enacting a no-fly or -sail zone—are more effective when enacted in conjunction with partners, the CNAS team factored this into the probability of success.

In addition to player actions, participants had to contend with randomly occurring events in the game. These events often involved actions by actors otherwise not represented in the game and introduced new dilemmas and decision-making constraints for the teams. Following each year-long turn, the CNAS team adjudicated and briefed out the results of player actions and the additional events in narrative fashion before the teams could plan their actions in response. This dynamic drove the game for a total of four moves, ending the game in 2029.

The game was run once, with a static set of players. While the participants represented a wide array of expertise, a single game produces a particular set of outcomes and therefore is limited to a single potential future. As games are not predictive, this single future is merely indicative. Nevertheless, there are significant insights that can be derived from this game. To make up for the limited iteration, CNAS researchers bolstered their insights with additional research, drawing on the real-world behaviors and actions of the three countries represented in the game.