IN touting his agency’s latest report, Superpowers and AGI, Jacob Stokes, deputy director of the Indo-Pacific Security Program at the Centre for a New American Security (CNAS), called during an online event on Sept 3 for the United States government to consider the feasibility of diplomatic, espionage, cyber, and kinetic measures to prevent China from achieving Artificial General Intelligence (AGI) first. 

CNAS is no ordinary American think tank. Established in 2007, the influential Washington, DC-based bipartisan institution specialises in US national security and defence policy. 

Heavily staffed by former high-ranking national security and intelligence officials such as Stokes from the Obama administration, Western media and policymakers frequently tout its “pragmatic, data-driven research” and praise its role as an ideological bridge. 

How seriously are its reports taken? 

CNAS reports regarding the perceived Chinese AI threat to the US are taken with utmost seriousness by federal policymakers, Congress, and executive officials. CNAS is not viewed as an outside commentator, but as an intellectual architect of modern American technology policy. 

The gravity with which Washington treats CNAS findings is evident in recent government policy actions, echoing frameworks laid out in seminal CNAS reports such as Red Lines, Adversarial Distillation, and Superpowers and AGI. 

This emerging discourse — openly contemplating preemptive military or kinetic action, such as targeting enemy data centers — reflects a dangerous escalation in how Washington frames the AI race. The debate has shifted from economic and technical competition to treating frontier AGI as an existential security threat akin to the nuclear parity concerns of the Cold War era. 

The American Sphere: Legislative and Policy Containment: In the US, bipartisan fear of losing technological leadership has forged a firm consensus around containment, reshoring, and aggressive export controls. 

Export Controls & Compute Blockades: The US Department of Commerce, through the Bureau of Industry and Security, continues to tighten restrictions on high-end AI accelerators (such as Nvidia and AMD chips), semiconductor manufacturing equipment (EUV/advanced DUV lithography), and High-Bandwidth Memory (HBM). 

Outbound Investment Screening: Executive orders and federal legislation restrict American venture capital and private equity from funding Chinese AI, quantum computing, and semiconductor ventures. 

Supply Chain De-risking: The CHIPS and Science Act channels tens of billions into domestic fab construction (e.g. Intel, TSMC Arizona, Samsung Texas) to reduce reliance on East Asian supply chains. 

Defense Integration & War-Gaming: Defense directives require integrating AI into autonomous weapons and command systems. Ideas from Washington think tanks regularly feed into Pentagon war games, where proposals ranging from offensive cyber-sabotage to kinetic strikes frame AI infrastructure as critical military targets. 

Multilateral Alliances: Under the "Small Yard, High Fence" doctrine, Washington coordinates with allies (including the EU, Japan, South Korea, the Netherlands and Taiwan) to enforce  aligned export regimes and control global AI standards. Malaysia and other Asean countries currently not committed to this doctrine are facing pressure on their China trade in chips and semiconductor manufacturing equipment. 

Beijing views US containment as an effort to suppress its sovereign development. In response, China has adopted a strategy centred on self-reliance, physical hardening, and asymmetric advantages. 

Full-Stack Self-Reliance ("Indigenisation"): Heavy state subsidies support domestic semiconductor efforts (such as Huawei’s Ascend ecosystem and SMIC foundries). Facing hardware bottlenecks, Chinese researchers focus on algorithmic optimisation, parameter-efficient fine-tuning, and open-source models to maximise performance from legacy hardware. 

Infrastructure Optimisation: Through the "Eastern Data, Western Computing" initiative, Beijing has centralised national computing hubs in energy-rich western provinces for maximum efficiency and grid resilience. 

Embodied AI & Industry: Recognising potential delays in pure software LLM races, China leverages its vast manufacturing base to lead in "embodied AI" — integrating intelligence into robotics, industrial automation, and smart logistics. 

Resource & Network Counter-Measures: China leverages its dominance over critical raw materials (gallium, germanium, antimony, rare earths) through export controls. Simultaneously, through the Belt and Road Initiative, it exports digital infrastructure across the Global South to build an alternative tech ecosystem. 

Physical Hardening: In response to Western discussions of targeting data centers, Beijing increasingly treats critical digital infrastructure as strategic defense assets — dispersing facilities geographically and fortifying them with air defense and cyber protocols. 

While the US strategy relies on compute denial, reshoring, and allied containment, China’s counter-strategy emphasises indigenisation, critical material leverage, and rapid industrial AI integration. 

The inclusion of kinetic strikes in mainstream American policy discourse signals that AI has crossed from a commercial domain into the core of strategic geopolitical priority and national security. Global cooperation on AI safety and development now appears increasingly out of reach. 

By normalising military contingencies around AI development, the current dynamic in the US opens two dangerous paths to catastrophe: a deliberate geopolitical miscalculation, or an over-automated command structure triggering runaway escalation. Both advance the Doomsday Clock towards a precipice no nation can afford to cross. 

Dr Lim Teck Ghee is an economic historian and policy analyst