September 2025

Rethinking a Political Approach to Nuclear Abolition
George Perkovich, Fumihiko Yoshida, and Michiru Nishida
Carnegie Endowment for International Peace, 2025

This book asserts that addressing the accelerating expansion of nuclear arms, growing nuclear risks and deepening geopolitical divisions requires strategies grounded in political-security realities more than the qualities or quantities of weapons. Notably, the authors acknowledge that their proposal may frustrate both disarmament and deterrence proponents. As they write, “In focusing on people and politics first, rather than the numbers and types of weapons and plans for their use, [… the book] will assume that preventing nuclear war and achieving nuclear disarmament are politically more difficult than the most avid proponents and opponents of nuclear weapons tend to recognize.”

Noting the difficulty of preventing nuclear war and further nuclear armament, the authors stress the “vital role of high-level political dialogue in advancing nuclear disarmament.” They say that the pragmatic approach is to acknowledge the “reality of nuclear deterrence while working to reduce both intentional and accidental risks.” The book outlines the unstable dynamics of the current nuclear landscape. It also explores three paths forward—abolition of nuclear weapons, “unstabilized” deterrence, or stabilized competition—and makes the case for the middle-ground option, stabilized competition, as the most pragmatic near-term solution “so long as politics preclude abolition.” Drawing on past arms control lessons, the authors articulate one large goal—to end nuclear overkill—and six guidelines to make deterrence more stable and accountable.

Despite little hope for disarmament, the authors argue that “it is possible that societies and some leaders will recognize the unnecessary danger of current trends and begin laying the groundwork for mutual restraints to 
be built when political changes allow more reasonable policies.”—SHIZUKA KURAMITSU

Beyond the Euromissile Crisis: Global Histories of Anti-Nuclear 
Activism in the Cold War
Luc-André Brunet and Eirini Karamouzi (eds.)
Berghahn Books, 2025

This open-access volume, funded by the UK Arts and Humanities Research Council, expands the scope of historical research on anti-nuclear activism during the Cold War beyond the well-known North Atlantic movements. The editors argue that the focus in existing literature on the Euromissile crisis of the 1980s and Intermediate-Range Nuclear Forces Treaty insufficiently represents the international reach of anti-nuclear movements in the late 20th century.

The edited volume collates 13 chapters and contributes to the field in four main ways: It broadens the geographical scope and includes regions that have been underrepresented in the discourse, such as Yugoslavia and French Polynesia; it highlights the connections and divisions across transnational movements, including disagreements on the definition of peace; it reassesses the established Cold War chronology by questioning the relevance of the Euromissile crisis for global movements; and it analyzes the relations between state and nonstate actors in the form of policymakers and peace movements. Overall, the book shows the diversity of movements, demonstrating that anti-nuclear activists followed no uniform practice in their campaigns.—LENA KROEPKE

September 2025
By Lt. Gen. John "Jack" N.T. Shanahan

As evidenced by the November 2024 agreement between Chinese President Xi Jinping and U.S. President Joe Biden, there is growing international consensus that artificial intelligence (AI) must never supplant human judgment in the authorization or execution of nuclear weapon launches.¹ Such momentous decisions should remain the sole domain of human leaders.

Although this nonbinding consensus is encouraging, it overshadows a more complex and less understood challenge, namely, that the integration of AI across the nuclear command and control/nuclear command, control, and communications (NC2/NC3) enterprise and beyond may create false confidence in the information that is shaping leaders’ situational awareness and influencing nuclear-related decisions.

Leaders of nuclear-weapon states must demand a thorough examination of the risks resulting from AI integration into all elements of the nuclear decision-making process. This analysis should encompass not just the traditional, well-documented components of the NC2/NC3 enterprise, but also all ancillary conventional platforms, sensors, intelligence systems, and information technology networks. Even seemingly unrelated applications, such as computer vision models deployed in Project Maven² or large language models (LLMs) used for problem scoping and mission analysis, can amplify the effects of other inputs that shape nuclear-related decisions.

Some incremental effects might not differ substantively from the pre-AI era, but the consequences of compounding failures in worst-case scenarios are potentially catastrophic. This reality underscores the need for a deep understanding of how AI affects every system, platform, and process that intersects with NC2/NC3. As no current modeling or simulation can capture the complexity of the entire NC2/NC3 enterprise during a real crisis, exercises, wargames, and tabletop simulations are crucial, yet insufficient. There may be no better case for applying the precautionary principle than when considering the role of AI in nuclear operations.

Definitions

Experts in nuclear policy, AI, and arms control have begun to analyze the opportunities, risks, and potential benefits of AI integration in the nuclear enterprise,³ but few studies offer a comprehensive assessment across the full spectrum of systems and networks that influence nuclear decision-making. Given AI’s still-nascent role in national security across nuclear-armed states, this gap reflects both the complexity of such analysis and the vast number of unknowns, to include many “unknown unknowns.” Nevertheless, the accelerating pace of AI integration underscores the urgency of such assessments.

The act of employing a nuclear weapon is often portrayed as a singular, dramatic decision made by the head of state during an escalating crisis or conflict. In reality, this decision is intended to be the culmination of a complex, all-domain process involving numerous interdependent systems. In some scenarios, such as launch-on-warning, timelines may compress to mere minutes. In others, leaders may have days or weeks to deliberate.

A critical first step in understanding this process is to distinguish clearly between NC2 and NC3, two terms often conflated under the broader “nuclear enterprise” umbrella. According to U.S. doctrine, NC2 is defined as the “exercise of authority and direction, through established command lines, over nuclear weapon operations by the President as the chief executive and head of state.”⁴ Its core requirements are that it must be assured, timely, secure, survivable, and enduring—delivering information and communications that enable presidential decision-making throughout a crisis.⁵ As in all U.S. military interpretations of command and control, the tenets of authority, direction, and control reside at the core of NC2.

NC2 comprises five mission-essential functions: force management, planning, situation monitoring, decision-making, and force direction. Force management refers to the assignment, training, deployment, maintenance, and logistic support of nuclear forces and weapons. Planning is defined as the development and modification of plans for the employment of nuclear weapons. Situation monitoring is defined as the collection, maintenance, assessment, and dissemination of information on forces and possible targets, emerging nuclear powers, and worldwide events of interest. Decision-making comprises the assessment, review, and consultation for employment of nuclear weapons. Force direction refers to the implementation of decisions regarding the execution, termination, destruction, and disablement of nuclear weapons.⁶

NC3, by contrast, refers to how those functions are executed: the means “through which Presidential authority is exercised and operational command and control of nuclear operations is conducted.”⁷ NC3 is part of the broader National Leadership Command Capability., It consists of the three broad mission areas of presidential and senior leader communications, NC3, and continuity of operations/continuity of government. NC3 encompasses facilities, equipment, communications, procedures, and people.

Together, NC2 (the ways) and NC3 (the means) are critical to the proper functioning of the U.S. National Military Command System (NMCS), which supports the desired ends: continuous, survivable, and secure nuclear command and control. The NMCS relies upon ground, airborne, maritime, space, and cyber systems to provide “unambiguous, reliable, accurate, timely, survivable, and enduring” warning about attacks on the United States, its allies, and its forces overseas.⁸

Despite these clear doctrinal definitions, even the most comprehensive descriptions of the NC2/NC3 enterprise fail to capture the full breadth of systems involved; the U.S. Air Force alone owns over 100 separate NC3 systems. They also neglect the staggering number of non-NC3 networks, platforms, sensors, and systems across all the military services that, in various ways, feed into, interface with, or could interface with the wider nuclear enterprise. This blurring of lines is even more pronounced in some other nuclear-armed states. Given the current trajectory, over the next decade AI is likely to be integrated not only into conventional military and intelligence platforms and networks, but also into NC3 systems. AI also is expected to play a larger role in supporting NC2 decision-making processes. A clear understanding of these definitions and their interconnections is essential to evaluating the implications of AI integration across the nuclear enterprise.

AI, NC2/NC3, and Strategic Stability

Scholars have argued that incorporating AI into NC2 processes and NC3 systems will introduce destabilizing dynamics among nuclear-armed states, particularly by diminishing direct human involvement in decisions about nuclear weapons use. Yet many of the same experts also recognize potential stabilizing effects in certain contexts. For example, by enhancing early warning and situational awareness, enabling more consistent detection and tracking of adversary systems, providing advanced tools for monitoring and verifying arms control and disarmament, and improving warhead and weapon safety, security, and reliability.

On balance, however, the risks to strategic stability⁹ from significantly accelerating nuclear decision timelines or reducing human involvement in launch decisions and execution are likely to outweigh the potential benefits.¹⁰

One especially destabilizing scenario involves the widespread application of AI to detect and continuously monitor the ground, sea, and airborne nuclear forces of other states. Adversary integration of AI-enhanced systems that can reliably locate mobile missile launchers or submerged ballistic missile submarines, which are considered the most survivable elements of a second-strike capability, could erode confidence in a state’s assured retaliation posture. This loss of confidence could in turn increase incentives for preemptive strikes, weakening the logic of deterrence and undermining strategic stability. The problem becomes even more acute if such surveillance and tracking technologies are paired with AI-enhanced automation of decision-making or launch processes, heightening first-strike incentives during a crisis.

At present, there is no evidence that any nuclear state, including the United States, intends to integrate AI directly in its NC3 systems or formal NC2 decision-making procedures. Nevertheless, in the absence of bilateral or multilateral agreements or established international norms circumscribing AI use, the likelihood that a state will take such steps will increase, while indirect integration into nuclear-adjacent systems is all but certain. Such integration may generate cascading effects that subtly and dangerously influence nuclear decisions, particularly if they remain undetected. Moreover, the mere perception that a state is integrating AI into its NC2/NC3 enterprise, or contemplating reduced human oversight over nuclear decisions, could itself be destabilizing. Given the uncertainty surrounding current and emerging AI capabilities, managing not only deployment but also perception will be a crucial component of strategic stability in the years ahead.

Indirect Pathways and Unpredictable Behaviors

The potential for AI to augment, accelerate, and automate tasks across the national security domain will lead inexorably to its expanded adoption. Although the expected benefits, such as greater speed, accuracy, effectiveness, and efficiency, may improve military operations, AI’s novel characteristics also introduce unique risks. These go well beyond automation of nuclear weapon launch decisions. In addition to the concerns already outlined, four compounding and interconnected areas demand urgent attention to avoid unintended or mistaken use of nuclear weapons.¹¹

First, the integration of AI across the NC2/NC3 enterprise introduces the risk of cascading effects and emergent behaviors. Errors resulting from flawed models or adversarial manipulation could propagate across interconnected systems, especially in the absence of sufficient human oversight.¹² Even seemingly minor downstream errors could produce disproportionate upstream consequences once multiple AI models are embedded across interconnected platforms and decision-support systems.

Compounding the danger is automation bias: the tendency to over-trust machines, particularly under crisis conditions marked by time compression, ambiguity, and extreme stress.¹³ These pressures could amplify the effects of unnoticed errors, producing tightly coupled feedback loops that distort the decision-making environment. Adversaries likely will exploit these vulnerabilities through AI-enhanced information operations designed to influence, disrupt, and corrupt the judgment of senior political and military leaders. For example, adversaries could use deepfake audio or video to fabricate crisis-related statements by senior political or military leaders or deploy LLMs to generate tailored disinformation campaigns that manipulate strategic warning indicators, public sentiment, or even internal military communications.

Emergent behavior poses a related concern. In AI, emergence refers to novel and unpredictable outcomes that arise from the interaction of relatively simple components—outcomes not explicitly programmed into the system.¹⁴ Ensembles of advanced AI models operating in dynamic, nonlinear systems may produce opaque and surprising results. Such unpredictability is particularly dangerous in nuclear operations, where the standard for accuracy and reliability is absolute. Former U.S. Navy Secretary Richard Danzig warned in Technology Roulette that “the introduction of complex, opaque, novel, and interactive technologies will produce accidents, emergent effects, and sabotage.” He cautioned that U.S. national security institutions risk losing control over their own creations.¹⁵ Although there is no evidence that any true self-learning AI systems have yet been fielded, developments are advancing rapidly, and the demonstration of emergent behaviors may only be a matter of time.¹⁶

Second, even if AI is monitored carefully within the formal NC2/NC3 enterprise, its use in adjacent systems—such as intelligence and surveillance sensors, conventional weapon systems, related information networks, and decision support systems—will indirectly influence nuclear decision-making. The same risks of cascading effects and automation bias apply, but with an added concern: Nuclear decision-makers may be unaware that AI is shaping their understanding of the operational environment. Comprehensive assessments of AI’s integration into conventional systems are essential, however complex those analyses may be. These AI-enabled assessments should map potential influences, ranging from most likely to most dangerous, and be shared with senior civilian and military leaders.

Third, large LLMs such as those powering ChatGPT or Perplexity may seem distant from the nuclear enterprise, yet arguably they pose a more immediate and underestimated risk. Generative AI (GenAI) tools are now being piloted across U.S. federal agencies. The Trump administration’s AI executive order calls for decisive action to ensure U.S. leadership in this field.¹⁷ Some departments and agencies have approved some models, tailored for government use, for experimentation and pilot projects. Wider adoption is inevitable.

Yet despite their rapid improvement, today’s open- and closed-source LLMs, to include the more advanced frontier and reasoning models, remain too brittle, skewed, and opaque to be trusted with mission-critical tasks like intelligence analysis for nuclear decision-making. Although this may seem obvious, there is a real risk that GenAI systems, used outside the NC2/NC3 enterprise, could influence assessments that ultimately shape nuclear decisions.¹⁸ At a minimum, U.S. national security leaders should require explicit disclosure of any use of GenAI in assessments that support nuclear course-of-action development. Even when GenAI is used for non-nuclear purposes, its outputs may shape broader assessments that feed into nuclear planning processes, sometimes without explicit attribution.

Finally, the integration of AI agents into commercial systems is advancing rapidly, and their eventual adoption in military operations and intelligence analysis is highly likely.¹⁹ These agents, capable of operating autonomously or semi-autonomously across systems and networks, pose enormous risks to nuclear operations. Their ability to take initiative, act without direct human intervention or even oversight, and learn over time creates scenarios in which behaviors become difficult to predict or control.

Cyber vulnerabilities and adversarial attack only compound the threat. As with cascading effects and emergent behavior, the risk posed by AI agents demands rigorous and detailed study. Mitigation must begin with the imposition of clear, human-defined constraints. With advanced AI, it will be far more difficult to enforce what agents must not be allowed to do than to direct them toward specific tasks.²⁰ Yet both types of constraints are essential if agents are to be integrated safely, if ever, within the NC2/NC3 ecosystem.

Circumscribing AI’s Nuclear Role

Although formal international agreements proscribing AI from authorizing or executing nuclear launches are unlikely in the near term, the 2024 Biden-Xi agreement and the emerging informal global consensus mark an important beginning. Still, these are only opening moves. It remains to be seen whether and how current or future administrations will follow through.²¹ These issues warrant candid discussion across Track 2, Track 1.5, and Track 1 dialogues, as well as in broader international forums. Depending on the state of relations between nations, these discussions could take place under the auspices of AI dialogues, arms control dialogues, or broader state-to-state dialogues that include an AI-nuclear nexus.

Leaders of nuclear-weapon states should issue clear public statements affirming the necessity of human oversight and control in all aspects of nuclear weapon deployment and employment. For example, as Gen. Anthony Cotton, the commander of U.S. Strategic Command, stated in his testimony to the Senate Armed Services Committee Strategic Forces Subcommittee in March 2025, “USSTRATCOM will use AI/ML to enable and accelerate decision-making.… AI will remain subordinate to the authority and accountability vested in humans.”²²

The potential benefits of AI for national security are significant, and its diffusion rate will only accelerate. The gravest risks of AI in the nuclear enterprise, such as automation of launch authority, are well understood and relatively easy to proscribe. But other risks are more subtle and systemic. The goal is not to halt AI adoption, but to deepen understanding of its unintended consequences, especially cascading effects, emergent behaviors, and hidden influences on intelligence and decision-making. Accompanying analyses, to include rigorous red teaming, should drive actions to mitigate the risks of indirect pathways and unpredictable behaviors while setting the stage for rapid human intervention as soon as unexpected or undesirable behaviors are detected.

In a real crisis, there may be no time to question how AI shaped the intelligence and analysis that informed nuclear recommendations. It is imperative, therefore, that senior military and intelligence leaders initiate detailed studies now, with the expectation that the studies will need continual refinement as AI continues to rapidly advance. These studies should be supported by experts within commercial technology companies and academic institutions and must address not only AI within NC2/NC3, but also within adjacent systems that may exert indirect influence. They also should assess the current state of human-machine integration, especially at the decision-support function, emphasizing the importance of the interfaces and interdependencies between humans and smart machines across the NC2/NC3 ecosystem.

In many instances, AI may add clear value with minimal risk. For instance, AI will be highly valuable in the behind-the-scenes work of warhead design, weapon safety and security, and scientific discovery. However, thorough evaluations might reveal instances where the potential cumulative effects of AI integration are so adverse that guardrails, other proactive actions, or outright prohibitions are necessary to preclude any possibility that a nuclear weapon is launched in error.

As AI continues to advance, it will match or exceed human performance across more national security functions. At the same time, humans are and will forever remain highly fallible. The juxtaposition of these two claims should not lead reflexively to the conclusion that AI is always the right answer. War is the ultimate human endeavor, and nuclear war its most consequential form. What sets humans apart—deductive and causal reasoning, contextual understanding, common sense, emotional awareness, and moral judgment—cannot be replicated by machines.

Ultimately, the decision to launch nuclear weapons must remain a distinctly human responsibility. National decision-makers, whether in the United States or any other nuclear- weapon state, must understand what role, if any, AI played in shaping the intelligence and recommendations presented to them. In that sense, the 2024 Biden-Xi agreement can be viewed not as the end of the debate, but as an inflection point marking the start of a global reckoning with AI’s role in nuclear stability.

ENDNOTES

1. The White House, Readout of President Joe Biden’s Meeting with President Xi Jinping of the People’s Republic of China, November 16, 2024. Beyond the Biden-Xi agreement, the broad international consensus remains an agreement in principle rather than in practice. For instance, the original draft of the State Department’s Political Declaration on Responsible Military Use of Artificial Intelligence and Autonomy contained language asserting that, as a best practice, governments “should maintain human control and involvement for all actions critical to informing and executing sovereign decisions concerning nuclear weapons employment.” This language does not appear in the final published version of the Declaration; in fact, there are no references at all to nuclear weapons or NC2/NC3. The Principles and responsible practices for Nuclear Weapon States, a 2022 U.N. working paper submitted by France, the United Kingdom, and the United States, states that “Consistent with long-standing policy, we will maintain human control and involvement for all actions critical to informing and executing sovereign decisions concerning nuclear weapons employment.” AI is not mentioned.

2. Project Maven is the name commonly used for the U.S. Department of Defense Algorithmic Warfare Cross-Functional Team, established in 2017.

3. See, for example, Johnson; Qi; Boulanin, et al.; Li; Kania; Depp and Scharre; Bajema; Schwartz and Horowitz; and Stokes, Kahl, Kendall-Taylor, and Locker:

James Johnson, AI and the Bomb: Nuclear Strategy and Risk in the Digital Age, Oxford University Press, 2023. Haotian Qi, “Chinese Perspective,” in Enhancing US-China Strategic Stability in an Era of Strategic Competition: US and Chinese Perspectives, edited by Patricia M. Kim, United States Institute of Peace, April 2021. Vincent Boulanin, Lora Saalman, Petr Topychkanov, Fei Su, and Moa Peldán Carlsson, “Artificial Intelligence, Strategic Stability and Nuclear Risk,” Stockholm International Peace Research Institute, June 2020. Xiang Li, “Artificial Intelligence and its Impact on Weaponization and Arms Control,” in The Impact of Artificial Intelligence on Strategic Stability and Nuclear Risk, edited by Lora Saalman, Stockholm International Peace Research Institute, October 2019. Elsa Kania, “The AI Titans’ Security Dilemma,” Hoover Institution, October 29, 2018. Michael Depp and Paul Scharre, “Artificial Intelligence and Nuclear Stability,” War on the Rocks, January 16, 2024. Natasha Bajema, “Will AI Steal Submarines’ Stealth?” IEEE Spectrum, July 16, 2022. Joshua Schwartz and Michael Horowitz, “Out of the Loop: How Dangerous is Weaponizing Automated Nuclear Systems?” arXiv, May 1, 2025. Jacob Stokes, Colin Kahl, Andrea Kendall-Taylor, and Nicholas Lokker, “Averting AI Armageddon: U.S.-China-Russia Rivalry at the Nexus of Nuclear Weapons and Artificial Intelligence,” Center for a New American Security, February 2025.

4. Deputy Assistant Secretary of Defense of Defense for Nuclear Matters, Nuclear Matters Handbook 2020 [Revised] (unofficial).

5. Ibid.

6. Department of the Air Force Curtis E. Lemay Center for Doctrine Development and Education, Air Force Doctrine Publication (AFPD) 3-72 Nuclear Operations, Nuclear Command, Control, and Communications, December 18, 2020.

7. Secretary of the Air Force Air Force Instruction 13-550, Air Force Nuclear Command, Control, and Communications (NC3), April 16, 2019.

8. Nuclear Matters Handbook 2020

9. For this paper, I use James Acton’s definition of strategic stability: “A deterrence relationship is stable if neither party has or perceives an incentive to change its force posture out of concern that an adversary might use nuclear weapons in a crisis.” I also accept his three broad interpretations of strategic stability: the absence of incentives to use nuclear weapons first (crisis stability) and the absence of incentives to build up a nuclear force (arms race stability); the absence of armed conflict between nuclear-armed states; and a regional or global security environment in which states enjoy peaceful and harmonious relations. (James Acton in Strategic Stability: Contending Interpretations, edited by Elbridge Colby and Michael Gerson, Carlisle, PA: Strategic Studies Institute, 2013, pp. 117-146.)

10. Colin Gray provides an alternative argument on the effects of technology on strategic stability. Although Gray did not address AI and the security dilemma, he analyzed the impacts of earlier technologies on strategic stability. He notes that strategic stability can result “when there is a rapid change in technological generations and considerable unpredictability concerning the building programs of rivals, yet where a tolerable balance of military power is maintained – albeit almost exclusively through competition.” My counter to Gray’s argument is based on AI as an unproven technology, especially in military applications: the high levels of uncertainty associated with AI’s actual performance, when juxtaposed against the inflated statements about AI emanating from leaders in China and the United States, suggest that the kind of “unrestrained competition” associated with well-understood and well-established naval competitions in the later decades of the 19th century will not lead to strategic stability in the AI competition (at least not for the immediate future). When it comes to AI, we still do not know how much we do not know. (Colin Gray, “Strategic Stability Reconsidered,” Daedalus, Fall 1980, 109(4), pp. 135-154.)

11. There are additional novel differences between traditional and AI-enhanced military systems that must be considered when assessing the possible effects of the introduction of AI across the NC2/NC3 enterprise. These include the control challenge, black box challenge, and the accountability challenge (I deliberately use “challenge” rather than “problem,” since all three must be treated as inherent, systemic features of all AI).

12. These are both examples of AI “corruption,” defined as the deliberate or unintentional manipulation of the data, hardware, or software of an AI-enabled system that causes the system to produce missing, inaccurate, or misleading results, to deny or degrade the use of the system, or to force the system to expose hidden information used in the training or configuration of the AI component. (National Academies of Sciences, Engineering, and Medicine, Test and Evaluation Challenges in Artificial Intelligence-Enabled Systems for the Department of the Air Force, Washington, DC: The National Academies Press, 2023.)

13. However, I am equally wary of human bias, manifested in the form of disregarding or dismissing the contributions of machines in favor of human judgment, heuristics, or gut instinct. In the nuclear context, these kinds of biases can be equally calamitous.

14. TedAI San Francisco, October 21-22, 2025.

15. Richard Danzig, “Technology Roulette: Managing Loss of Control as Many Militaries Pursue Technological Superiority,” Center for a New American Security, June 2018.

16. Although LLMs demonstrate behavior that can be construed as “continual learning,” they are not true online learning systems. The debate continues as to whether they are stochastic parrots or something more advanced. See, for example, Steffen Koch, “Babbling Stochastic Parrots? A Kripkean Argument for Reference in Large Language Models,” Philosophy of AI, Vol. 1, 2025.

17. The White House, “Removing Barriers to American Leadership in Artificial Intelligence,” January 23, 2025. Examples of approved models include Claude Gov, Defense Llama, ChatGPT, NIPRGPT, and CamoGPT.

18. There have been only two instances of nuclear weapon use in conflict and relatively few global crises in which their use may have been seriously considered. As a result, the volume of real-world data available to train LLMs on nuclear decision-making is extremely limited. Although data from war games, simulations, experiments, and decades of nuclear deterrence and escalation theory will inform such models, any GenAI-derived responses to prompts related to nuclear weapons employment should be interpreted with great caution. For the foreseeable future, the human expert-smart machine combination will remain the only solution.

19. An AI agent is a software-based system capable of perceiving its environment, making decisions, and acting to achieve specified goals. They may combine sensing, reasoning, and learning capabilities to operate autonomously or semi-autonomously.

20. In the nuclear enterprise, the combination of emergent behavior and misalignment, in which LLMs and agents collaborate to achieve objectives beyond or contrary to human-defined boundaries, will be highly problematic. See, for example, Miles Wang, Tom Dupré la Tour, Olivia Watkins, Alex Makelov, Ryan A. Chi, Samuel Miserendino, Johannes Heidecke, Tejal Patwardhan, and Dan Mossing, “Persona Features Control Emergent Misalignment,” arXiv, June 2025.

21. The 2025 White House AI Action Plan and accompanying executive orders do not address the potential for follow-on AI dialogues between the United States and China, or with any other state.

22. Testimony of Gen. Anthony J. Cotton before the Senate Armed Services Committee Subcommittee on Strategic Forces, March 26, 2025.

September 2025
By William C. Potter and Sarah Bidgood

Some meetings of nuclear Nonproliferation Treaty (NPT) member states are remembered for artistic performances as well as nonproliferation and disarmament achievements.

They include the remarkable collection of cartoons created at the 1995 NPT Review and Extension Conference by U.S. delegate Carl Stoiber; the 2002 NPT preparatory committee chair’s “factual summary,” sung to the tune of the Beatles’ “Yesterday” by Swedish diplomat and former pop star Henrik Salander; and the poorly executed standup comic routines that masqueraded as “rights of reply” at the end of many NPT sessions since the 2018 preparatory committee meeting in Geneva. In terms of a theatrical metaphor for the NPT review process, perhaps the most appropriate is Samuel Beckett’s classic play, “Waiting for Godot.”

Just as it is hard to pin a definitive label on Beckett’s play—existential drama, theater of the absurd, or tragicomedy—the NPT review process looks very different to key stakeholders. As with the play, the current review cycle, which culminates next April at the 2026 NPT Review Conference, is struggling with existential issues. They involve complex power dynamics among unequal parties, language that tends to obscure rather than clarify, a cyclical process in which the key players often repeat the same arguments, and boundless faith in an elusive promise.

Faced with these challenges, it should have been little surprise that the third session of the current NPT review process cycle held April 28-May 9 failed to agree on recommendations to present to the 2026 Review Conference. If that conference also fails, it could further erode the NPT and nonproliferation efforts in general.

A Ritual of Repetition

After the NPT’s entry into force in 1970, representatives from most of the treaty’s 191 member states have gathered for four weeks at review conferences to take stock of the operation of the treaty. Except for during the COVID-19 pandemic, the conferences have been held every five years. In between review conferences, preparatory committee meetings typically are held for two weeks in each of the three years prior to the next review conference. In order to make the review process more results-oriented, the Final Document of the 2000 Review Conference mandated that the third session of the review cycle should strive to produce consensus recommendations for the upcoming conference.

In fact, this mandate has never been fulfilled, and the 2025 preparatory committee meeting was no exception. Instead, delegates from 145 states gathered in stuffy conference rooms at UN headquarters in New York, dutifully reiterated their support for the NPT’s three pillars, read aloud their well-known statements during the general debate and concluded the meeting without any agreed recommendations for the 2026 Review Conference. This outcome paints a gloomy prognosis of what might be expected in 2026.

The 2025 preparatory committee, chaired by Harold Agyeman, Ghana’s UN ambassador, began on a promising note as UN High Representative for Disarmament Affairs Izumi Nakamitsu reminded delegates of the heightened nuclear risks confronting the world and the need for the NPT to evolve to meet them. Luck, she argued, is not a strategy, and one cannot afford to engage only with the challenges of yesterday. In addition, she observed how many states felt frustration that “nuclear disarmament was always viewed as something that would come later, when conditions were better.”¹

Nakamitsu also called for states to “pay more attention to the intersection between nuclear weapons, new technologies, and new domains of conflict” and highlighted the link between the inalienable right to peaceful nuclear use and sustainable development. Moreover, she tellingly urged states to “be flexible and willing to compromise, and above all, engage in concrete and practical preparation.”²

Regrettably, her wise words appeared to fall mainly on deaf ears until the morning session of the last day when, at last, a meaningful interactive debate took place on strengthening the review process. This debate, facilitated by Agyeman, was the only closed session of the preparatory committee meeting and mainly revolved around the issues of transparency and accountability.³ The discussion, which continued over an unusual luncheon meeting to which all interested delegates were invited, led to a significant convergence of views, but consensus still proved impossible, largely due to China’s opposition over the issue of reporting. On this matter, the Chinese delegation was adamant that allies of nuclear-weapon states enjoying “nuclear umbrella” status also must report on their extended deterrence arrangements.⁴

In stark contrast to China’s assertive stance on most issues before the preparatory committee—often linked by the head of the delegation (who this year, unusually, was not of ambassadorial rank) to the issue of “nuclear sharing”—the United States was exceptionally restrained. This likely was due to the lack of senior arms control and nonproliferation leadership in the U.S. Department of State at the start of the second Trump administration and the corresponding absence of instructions. As a consequence, there was more continuity than change in U.S. positions on most issues. One exception was the more restrained U.S. interventions regarding Russia’s invasion of Ukraine; the dangers posed by the Kremlin’s attacks on, and occupation of, Ukrainian civilian nuclear facilities; and bellicose rhetoric that eroded norms against nuclear weapons use.

It was not only the United States, however, that had little appetite to discuss the erosion of the nuclear taboo, the prospect of future proliferation, and possible state withdrawals from the NPT. None of the nuclear-weapon states were prepared to pay even lip service to the grave humanitarian consequences of nuclear weapons, much less the benefits of the Treaty on the Prohibition of Nuclear Weapons (TPNW). In addition, few states from any of the regional and political groupings, north or south, were ready to acknowledge the need to discuss concrete proposals related to the risks posed to all three NPT pillars by disruptive technologies, including artificial intelligence. Most states also were reluctant to focus on the importance of states adhering to past review conference obligations such as the action plan from 2010, the last review conference to produce a consensus final document.

Although the nuclear-weapon states deserve much of the blame for the collective amnesia about prior disarmament commitments, many non-nuclear-weapon states were willing to defer to the demands of those nations that regard themselves as responsible nuclear-weapon states. This acquiescence was most vividly on display when a number of disarmament “white knights” accepted the argument that an innocuous draft statement on disarmament and nonproliferation education, which initially enjoyed widespread support, was a TPNW stalking horse.⁵

To his credit, Agyeman circulated his “Draft Chair’s Elements Paper on Strengthening the Review Process” shortly before the start of the preparatory committee meeting. He also sought to build support for his recommendations by omitting from his subsequent draft papers many of the issues that had proved especially contentious in the past. These initiatives by the chair contributed to a less toxic atmosphere than was the case in 2023 and 2024. Nevertheless, he chose to conclude the preparatory committee meeting before its time expired, resigning himself to the fact that consensus recommendations were beyond reach. This led him to submit a working paper under his own authority. The paper still has not been publicly released, although presumably it corresponds closely to the recommendations in the “Revised Draft Recommendations to the Review Conference of the NPT” that was submitted on the penultimate day of the meeting.⁶ Therefore, as was the case in most prior third sessions of the review process cycle, the 2025 Preparatory Committee meeting ended with a whimper rather than a bang.

On to the Next Review Conference

Based on the outcome of this year’s meeting, it is hard to be optimistic about what awaits NPT states-parties next April. To be sure, many delegations will continue to express hope that progress on disarmament and nonproliferation is still possible, although others may conclude that the review process has become an end in itself. Yet it is also possible that the NPT states will be hard-pressed to agree on specific steps to achieve these objectives, or on language to characterize the progress they have made during the past review cycle.

Against a backdrop of multiple setbacks to the international nonproliferation regime, including multiple attacks on nuclear facilities in conflict zones, renewed interest among some ostensibly friendly states in proliferating, the absence of effective regional and political coalitions to bridge major divides among NPT stakeholders, and continuing toxic atmospherics in the review process, it is hard to imagine a 2026 NPT Review Conference outcome that most states-parties would regard as successful.

Israeli and U.S. Attacks on Iran’s Nuclear Facilities

Although Russia’s full-scale war against Ukraine and its attendant nuclear risks are likely to remain a contentious issue at the upcoming review conference, the June 2025 airstrikes by Israel and the United States against Iranian nuclear facilities are likely to share the spotlight. The precise impact of the strikes on the conference deliberations remains to be seen, but they already have precipitated renewed and more credible threats from Tehran to leave the treaty if international sanctions are snapped back into place.⁷ Should Iran act on these threats, either in the leadup to or at the review conference, it would exacerbate fissures that already threaten to derail conference negotiations.

Regardless, such attacks by a nuclear-weapon state on a non-nuclear-weapon state’s nuclear facilities could well become the issue over which the conference collapses. A majority of states likely will push for inclusion of language in any outcome document condemning the Israeli and U.S. attacks, which the U.S. delegation surely will oppose. Unless compromise language can be adopted to reconcile these views, which seems unlikely, the result could be a showdown similar the last review conference in 2022. There, the Russian delegation blocked consensus in large part because of objections to language referencing Russia’s occupation of the Zaporizhzhia nuclear power plant in Ukraine and the nuclear safety crisis that ensued.⁸

Expiration of New START

Another looming development that will complicate the tasks of review conference negotiators is the expiration of the New Strategic Arms Reduction Treaty in February 2026. Although this agreement is no longer being implemented following Russia’s suspension of participation in 2023, it remains the only bilateral arms control treaty still legally in place between Russia and the United States and is an important symbol of their disarmament commitments. If negotiations on a follow-on treaty are not in sight by the start of the review conference, many non-nuclear-weapon states may demand that the final document call on Russia and the United States to return to the bargaining table.⁹ Support for this measure would grow significantly should Moscow or Washington be perceived as preparing to resume nuclear testing.

Forging a consensus outcome also is apt to be impeded by the absence of obvious candidates among individual treaty members or regional and political groupings who effectively can bridge major divides among key stakeholders. Among the topics that will require creative bridge-building are the implementation of past review conference obligations, the relative urgency for concrete steps toward nuclear disarmament and nuclear risk reduction, the prospect and consequences of new nuclear proliferators, the lack of headway toward regional security in the Middle East and the Korean peninsula, and the importance of transparency and accountability in the review process.

A small group of states called the New Agenda Coalition had considerable success in mediating contentious nuclear disarmament issues in the past, notably at the 2000 Review Conference, but there are no obvious coalitions that can perform a similar function today. The lack of strong leadership and the fragmentation within traditional political groupings, including the Non-Aligned Movement, also reduces the ability of these bodies to display flexibility in negotiations and to bring their members on board in support of proposed compromise language. In recent years, a possible exception to this general tendency was the ability of countries from Latin America and the Caribbean to display considerable solidarity in promoting the disarmament interests of many non-nuclear-weapon states, but the cohesion of this regional grouping is likely to be diminished at the next review conference due to political changes and increased friction involving some key members.

Atmospherics

Identifying strategies to overcome these challenges would be difficult under the best of circumstances, but is likely to be made more so by the lack of respect, trust, flexibility, and empathy among key NPT stakeholders. Perhaps nowhere is this more apparent than in the toxic practice of delivering acrimonious and often unscripted rights of reply at many NPT sessions, a ritual that China, Iran, Russia, and the United States appear unlikely to abandon.

To some extent, the tenor of debate at NPT meetings is set by the heads of key national delegations and their overseers in capitals. It is hard to anticipate what the picture will look like at the upcoming review conference based on the low-level representation at the 2025 preparatory committee meeting on the part of China, Russia, and the United States; none had an ambassador present. The United States will almost certainly have a more senior head of delegation, possibly at the undersecretary level, who will advocate positions more closely attuned to the assertive ones being espoused by U.S. diplomats in other international fora. This is likely to take the form of very hostile U.S. interventions against all Chinese positions, a stance that will be met by equally aggressive anti-U.S. rhetoric from China.

It is less clear how vocal the United States will be with respect to Russian aggression in Ukraine and other Russia-led initiatives related to NATO, nuclear sharing, the Middle East, and North Korea. Conceivably, the United States will let other NATO members take the lead in pushing back on at least some of these issues.

Perhaps the wildcard issue with the greatest potential to disrupt the review conference and provoke widespread anti-U.S. interventions is U.S. opposition to all efforts to reference UN sustainable development goals, even if they relate to almost universally shared support for peaceful nuclear use. This issue has not been the subject of controversy at prior NPT meetings, but should it arise at the review conference, it could well doom any prospect of a consensus final document.

Potential Areas for Progress

If this non-consensus scenario materializes, then the final act of this NPT review cycle promises more of the same unsatisfying saga. Existing power disparities among nuclear- and non-nuclear-weapon states will only become more apparent and longstanding conflicts more entrenched, with little reason to believe the next review cycle will be any more successful than the two previous ones. Notwithstanding this gloomy prognosis, one can identify a few positive nuclear developments that the review conference might note and other areas where there may be potential for a significant convergence of views, even if it falls short of consensus.

One promising development is the decision in June by Syria’s new government to allow inspectors from the International Atomic Energy Agency to visit suspected nuclear sites. IAEA Director-General Rafael Mariano Grossi has said that this action provides an opportunity to close a chapter of Syria’s past and “bring greater clarity” about its prior nuclear activities.¹⁰ The agency’s closure of its Syria “nuclear file” would remove a longstanding, contentious issue from the review process agenda.

Probably the greatest convergence of views among NPT members regarding core elements of the treaty pertains to the use of nuclear technology for nonmilitary purposes. Support for the “inalienable right” and access to peaceful use enjoys widespread support among nuclear- and non-nuclear-weapon states. Notwithstanding the possibility of controversy should some states seek to link an “inalienable right” to peaceful uses to UN sustainable development goals, it is conceivable that the review conference could carve out language that would enjoy consensus support.

Another area where near consensus may be possible concerns strengthening the NPT review process. Considerable progress was made at a special working group on this topic that met for one week in Vienna immediately before the 2023 preparatory committee meeting.¹¹ Although Russia ultimately made it impossible to reach any consensus recommendations, it is conceivable that some creative formulation addressing the need for increased transparency and accountably could be found to overcome likely resistance by China and Russia at the 2026 Review Conference. This year’s preparatory committee chair suggested several possible approaches in his “Revised Draft Decision: Strengthening the Review Process.”¹² That said, although transparency and reporting are important to many states, they are fundamentally procedural matters that do not address the most significant shortcomings of the nonproliferation regime.

Waiting for Consensus

Diplomats use language to both obscure and clarify, a practice that is evident in the annals of the NPT review process. Two words that merit special attention in this context are “success” and “consensus.” What, in fact, is meant by a “successful” review conference? Does it depend on the negotiation of a consensus outcome document? Should one equate consensus with unanimity or the absence of the need for a vote?

Adopting a final document without a vote has been the traditional benchmark by which NPT review conference success has been judged by diplomats and outside analysts. Yet, it is a dubious metric by which to assess the conference’s impact on the nonproliferation regime, especially considering the frequent failure of states to adhere to the consensus documents they negotiated.

In addition, a consensus outcome may indicate widespread and enduring support for disarmament and nonproliferation principles and actions, or it may simply signify acceptance of recommendations based on a very low common denominator. Importantly, fixation on the tradition of consensus decision-making ignores the fact that voting on substantive and procedural matters is explicitly permitted in the rules of procedure adopted at the first NPT Review Conference in 1975.¹³

The need to explore alternatives to consensus-based outcomes is not new, although most states are reticent to endorse the idea. The most detailed discussion of alternative modalities was produced by veteran U.S. State Department official Robert Einhorn ahead of the planned 2020 Review Conference. These proposals include adopting an entire document by consensus but resorting when necessary to formulations that reflect differences between parties (for example, “‘some believe’ X, while ‘others believe’ Y.”)¹⁴ He also suggests a final document that includes both consensus and non-consensus components, such as a “declaration issued at a very senior level … reaffirming the parties’ strong support for the treaty,” a series of recommendations that enjoy consensus, and a list of those that do not.¹⁵

In consultations with states-parties over the next nine months, the president-designate of the 2026 Review Conference, who is expected to be from Vietnam, could probe which of these or other alternative outcomes might be acceptable to states if it becomes clear that traditional consensus, for a third consecutive time, is likely to be out of reach.¹⁶

Critics have not been kind in their assessment of recent NPT review conferences, although some of the leading protagonists, such as the president of the long-delayed 2022 conference, received high marks for their performances. As the 2026 conference approaches, the challenge before states-parties is not whether they can reach a consensus final document regardless of its content. Rather, it is whether delegates can respond with practical recommendations that are commensurate with increasingly acute existential threats.

As with Beckett’s characters, NPT delegates appear content to pass their time tethered to routines and practices they cannot abandon and conversations they cannot resolve. It remains to be seen if the actors on the NPT stage still believe in the original treaty script. If not, the NPT review process and the 2026 Review Conference are in danger of becoming irrelevant.

ENDNOTES

1. Notes taken by one of the authors April 28, 2025, at the opening session of the third session of the preparatory committee for the 2026 Review Conference of Parties to the Treaty on the Non-Proliferation of Nuclear Weapons.

2. Ibid.

3. The irony of having a debate about transparency in a session that was closed to civil society was not lost on many of the delegates, including the French UN ambassador, who spoke forcefully but unsuccessfully against the procedure.

4. Commentary based on the notes taken at the closed session by one of the authors.

5. Personal observations of one of the authors.

6. Revised Draft Recommendations to the Review Conference of the NPT, NPT/CONF.2026.PC.III.CRP.4/Rev.1, May 8, 2025.

7. Yashraj Sharma, “What is the NPT, and Why Has Iran Threatened to Pull Out of the Treaty?” Al Jazeera, June 17, 2025; Tim O’Connor, “Iran Considers Leaving Nuclear Treaty if Snapback Sanctions Triggered,” Newsweek, July 23, 2025.

8. Gabriela Rosa Hernandez and Daryl Kimball, “Russia Blocks NPT Conference Consensus Over Ukraine,” Arms Control Today, September 2022.

9. Joint Statement on New START delivered by Austria on behalf of 24 states on May 9, 2025. In comments to reporters on July 25, 2025, U.S. President Donald Trump called for maintaining the limits put in place by New Strategic Arms Reduction Treaty, observing, “That’s not an agreement you want expiring.” See “Trump says he wants to maintain nuclear limits with Russia,” Reuters, July 25, 2025.

10. Gavin Blackburn, “Syria to Give UN Inspectors Immediate Access to Former Nuclear Sites, IAEA says,” Euronews, May 6, 2025.

11. William C. Potter, “Behind the Scenes: How Not to Negotiate an Enhanced NPT Review Process,” Arms Control Today, October 2023.

12. Revised Draft Recommendations to the Review Conference of the NPT, NPT/CONF.2026.PC.III.CRP.4/Rev.1, May 8, 2025.

13. Surprisingly, few NPT diplomats or scholars appear to be aware that before the start of the 1995 NPT Review and Extension Conference, intense deliberations took place among delegates about various contingencies for voting on the issue of indefinite extension of treaty, the outcome of which was by no means certain. At the time, much of the debate was not about the wisdom of voting but rather about the modalities that would be involved, and especially if a vote should be open or secret.

14. Einhorn cautions against an overreliance on this approach, noting that it would “highlight divisions among the parties rather than convey the impression that the parties were basically in agreement.” Robert Einhorn, “The 2020 NPT Review Conference: Prepare for Plan B” (Geneva, Switzerland: UN Institute for Disarmament Research, 2020): p. 17.

15. Robert Einhorn, “The 2020 NPT Review Conference: Prepare for Plan B,” UN Institute for Disarmament Research, 2020: p. 18.

16. At the time of this writing, the Non-Aligned Movement had yet to formally nominate Vietnam as its candidate for the NPT presidency, but it is expected that the Vietnamese Permanent Representative in New York will serve in that capacity.

William C. Potter, the founding director of the James Martin Center for Nonproliferation Studies, is the Sam Nunn and Richard Lugar professor of nonproliferation studies at the Middlebury Institute of International Studies at Monterey. Sarah Bidgood is a postdoctoral fellow in technology and international security at the University of California Institute on Global Conflict and Cooperation and a former director of the Eurasia nonproliferation program at the James Martin Center for Nonproliferation Studies.

September 2025
 

The Logic of Strategic Substitution

Under the Nuclear Shadow
By Fiona Cunningham
Princeton University Press, 2025

Reviewed by Elsa B. Kania 

Under the Nuclear Shadow is a critical contribution to understanding the rationales informing China’s military force posture and evolving approach to deterrence. Author Fiona Cunningham introduces a theory of strategic substitution and uses it to examine the logic behind China’s drive to develop information-age capabilities. As the Chinese People’s Liberation Army (PLA) faces the limited war dilemma of defining the challenge of achieving objectives through military force while mitigating risks of nuclear conflict, its pursuit of coercive leverage has motivated significant investments and transformation of its force posture. In particular, Chinese leaders and military strategists have regarded the development of information-age capabilities, especially in advancing cyberspace operations, counterspace weaponry, and precision conventional missiles, as uniquely advantageous, given their greater flexibility and potential for improved credibility. The strategic rationales and historical perspectives that this book provides are invaluable to our understanding of the contours of Chinese military power.

China’s Search for Leverage

The book traces the trajectory of China’s military modernization and examines several critical inflection points in which “leverage deficits” provided an impetus for Chinese leaders to prioritize pursuing certain information-age capabilities. In particular, the 1995-1996 Taiwan Strait crisis, when the U.S. military deployed aircraft carriers off the coast of Taiwan in response to China’s attempted coercion, and the 1999 Belgrade embassy bombing, which Beijing believed to be deliberate, can be regarded in retrospect as moments when Chinese leaders became acutely aware of shortfalls in their capabilities relative to the United States. These realizations created catalytic impacts driving Beijing’s decision to develop capabilities that Washington, regarded as a powerful potential adversary, could fear. While Cunningham characterizes these moments primarily as political crises, the military motivations and considerations are closely entwined, especially after factoring in the singular demonstrations of U.S. military capability in each instance.

The theory of strategic substitution is presented as a puzzle and a unique choice by China to pursue an alternative approach relative to adopting a nuclear-first-use posture (brinksmanship) or to develop superior conventional capabilities. By contrast, China decided to bet on information-age capabilities, here defined as cyberspace, counterspace, and conventional precision missile capabilities. This approach could be regarded as highly risky yet pragmatic; Chinese leaders were willing to bet on new domains and more untested technologies, yet also regarded these specific capabilities as quicker, feasible options for accelerated development relative to more conventional alternatives.¹

This book provides a deeply researched and authoritative accounting of how each of these elements of force posture evolved. Relative to nuclear weapons, which Chinese leader Mao Zedong once dismissed as “paper tigers,” given the perceived limitations of the credibility of nuclear threats, precision strike, counterspace, and cyberspace capabilities are regarded as more usable, flexible, and credible. As such, these capabilities can provide the “coercive leverage” that China seeks, while allowing for tight control of the risks of escalation. Cunningham defines the posture that follows from this capability development as one of “calibrated escalation” that provides a range of options along a ladder up to but still “under the nuclear shadow.”

Ultimately, these capabilities, which might be characterized as a new triad complementing the fully nuclear triad that China has been developing in parallel, can be regarded as instruments of an overall system of strategic deterrence that is extending increasingly beyond China’s nuclear arsenal to incorporate new concepts of information deterrence, as well as cyber and space deterrence. Pursuant to the increased importance of new domains and emerging capabilities, from artificial intelligence to deep sea developments, the factors that may influence deterrence are increasingly complex and multifaceted.

While Cunningham looks back to the 1990s for moments with catalytic impacts on China’s force posture, more recent noteworthy events may be relevant in anticipating future developments. The reported misperceptions among Chinese leaders that caused Chairman of the Joint Chiefs of Staff Gen. Mark Milley to engage his counterpart in the PLA Joint Staff Department in October 2020 and January 2021 in an effort to provide assurance that there was no U.S. intention to attack China may be regarded as another episode with potential catalytic impacts that illustrates the recent mistrust in China-U.S. relations.² The PLA’s response to House Speaker Nancy Pelosi’s visit to Taiwan in August 2022 also may have proven a moment that demonstrated a proof of concept for applying military power in a retaliatory manner and for the exercise of pressure, a trend that has become more prominent in the years since then.

This force posture and strategic outlook possess certain limitations. For China, the use of such information-age capabilities in signaling for coercive or deterrent purposes is inherently challenging. Among the mechanisms that Chinese military strategists anticipate using are deliberate disclosures, as by testing and demonstrations that publicly showcase such capabilities. The risks of entanglement in the underlying command systems and the dependencies between these domains create complexities, which may complicate the carefully modulated approach to escalation that Chinese leaders envision enacting. In particular, the increasing nuclear-conventional entanglement that has become a troubling feature of China’s overall posture has raised concerns about risks of escalation.³

Beijing’s capacity to convey the intentions behind certain activities and to influence potential adversaries’ perceptions of these capabilities constitutes another difficulty where there is significant potential for misperception. For instance, although the book mentions Volt Typhoon in passing, this is an episode that merits more detailed examination and consideration as an illustration of China’s approach. In particular, the reported pre-positioning of malware in U.S. and allied critical systems and infrastructure appears to highlight the realization operationally of certain concepts that Chinese military writers had long discussed conceptually.⁴ Such peacetime operations are also consistent with a view of the cyber domain that regards a blurring of boundaries between peace and warfare, as well as the inherent interrelatedness of reconnaissance and offensive operations, as features of this new domain of conflict.

In that particular episode, even more curious was the apparent acknowledgment by Chinese officials of responsibility for these attacks and their statements to reportedly surprised U.S. counterparts that the purpose behind this operation was to respond to U.S. support and assistance for Taiwan.⁵ Unspecified in public reporting is whether that activity should be regarded as a punitive, coercive, or deterrent measure. Absent such explicit conveying of intentions, the intent behind cyber operations can be easily misconstrued; when the boundary between preparation for and imminent conduct of offensive operation is inherently uncertain from a technical perspective, presence on sensitive systems and networks, especially in a moment of crisis, could be taken as an indication of aggressive intentions, rather than a deterrent measure.

Questions of Command

Among the notable avenues for future inquiry is how China’s approach to command and control influences its capacity to leverage these information-age capabilities. Cunningham traces the evolution of PLA command of cyberspace capabilities from an earlier more freewheeling environment to increasingly centralized control, which prompted creation of the Strategic Support Force and later the PLA Cyberspace Force. The inherent question is whether Chinese leaders can achieve the desired strategic effects while limiting the delegation of control over strategic capabilities when the Chinese Communist Party prefers to maintain more centralized control.

To date, China’s capacity for coordination appears limited. Cunningham notes that there is little evidence of a structure to coordinate actions associated with strategic deterrence. Although there is almost certain to remain a high degree of opacity associated with these processes, the Central Military Commission’s Joint Operations Command Center presents a feasible mechanism for such coordination. The reported construction of a new potential wartime command center in Beijing, seemingly buried deep enough underground to withstand a nuclear attack, highlights the apparent intensity of concerns driving these developments.⁶ Meanwhile, Chinese political leaders may recognize the necessity of more delegated mechanisms of command in certain domains and for certain missions that place greater premium on the exercise of initiative.

The Nuclear Posture Puzzle

Among the inherent challenges associated with such a book is the dynamism of the topic. Under the Nuclear Shadow has been published against the backdrop of an active debate on the dramatic expansion of China’s nuclear arsenal. At first glance, the extent of China’s investment in modernizing, diversifying, and expanding its nuclear forces—which the U.S. Defense Department’s 2024 “China Military Power” report predicts will probably continue over the next decade⁷—would appear to contradict some of the premises in Under the Nuclear Shadow. Cunningham concentrates on the drivers for China’s development of non-nuclear strategic capabilities to substitute for nuclear threats. Yet the logic underlying these developments parallels that of strategic substitution, insofar as the PLA’s pursuit of more diversified nuclear capabilities allows greater optionality.

In this regard, the trend overall has been to expand options within and along the nuclear section of the escalation ladder. This drive for nuclear overmatch and escalation dominance appears to reflect a juxtaposition of a degree of insecurity, including in the context of U.S. missile defense capabilities, and the extent of China’s military ambitions, with the drive to become a world-class military by mid century.

To some extent, there may be certain tensions or tradeoffs from a resourcing perspective between the logics of strategic substitution and the pursuit of nuclear overmatch. In parallel, there are indications that the PLA is increasing the readiness of its nuclear forces with a focus on rapid response, which shifts closer to a launch on warning posture. While there are reasons to hope that an approach of calibrated escalation, such as Cunningham characterizes China’s approach, will prevail, grounds for concern also arise about the apparent confidence of Chinese leadership that escalation can be controlled, especially given the potential for mistakes in signaling or misperception.

Author’s Contributions

Overall, Cunningham’s research provides important theoretical contributions to understanding Chinese military power. Her characterization of China’s force posture as one of calibrated escalation also possesses utility and value in projecting how Beijing might employ these capabilities for coercion in conflict. The theoretical framework of strategic substitution accounts for the logic of China’s bets on information-age capabilities as an attempt to offset U.S. military advantages.⁸ This drive for leverage centered on achieving quick, credible capabilities within a shorter time for the three elements of force posture is worth noting. However, in parallel, Chinese military modernization has pursued capabilities with longer timelines for development.

Perhaps the most fundamental consideration raised by this book is how to project the future trajectory of Chinese military power. The author articulates three plausible scenarios for China’s force posture: continuation of strategic substitution, change to a nuclear-first-use posture, or pursuit of conventional victory. These options appear far from mutually exclusive. At present, China appears to be pursuing a combination of the three trajectories, as reflected in continued development of information-age capabilities juxtaposed with dramatic expansion of its nuclear posture—albeit without a clear change to its position on nuclear first use—and significant expansion of conventional capabilities. If Chinese leaders start to face resource constraints because of economic headwinds, the telling sign may be how resources are prioritized and allocated among these different categories of capabilities.

This framework of strategic substitution provides a useful approach that has potential applicability to both nuclear and non-nuclear states. At a time when U.S. security guarantees and extended deterrence commitments are starting to be called into question, there are rationales for some U.S. allies and partners to pursue more sources of leverage and independent capabilities. For instance, Taiwan is pursuing new asymmetric capabilities through foreign military sales and domestic procurement, such as precision strike missiles, including coastal defense cruise missiles, and a growing number of unmanned systems.⁹ Relative to when China initially embarked on a quest to develop information-age capabilities, its tendency today toward commercial developments may make these strategies and the development of comparable capabilities more accessible to a greater range of states.

Ultimately, this work raises critical questions as the Chinese military looks to the future of power and warfare. Implicit in the theory of strategic substitution is how Chinese political and military leaders make decisions about the bets they place on technologies under conditions of uncertainty. To no small extent, China’s pursuit of information capabilities also occurred against the backdrop of concepts of a revolution in military affairs, the notion of a dramatic transformation that centers on the decisive impacts of information in warfare. These ideational influences have shaped decision-making on which technologies to pursue and prioritize.

Among the potential avenues for future research are how China’s drive for military innovation, which is producing new, yet untested capabilities, might be integrated into its overall force posture. Beyond the information-age capabilities that Cunningham characterizes as critical to deterrence, Chinese military strategists also are starting to explore the impacts of unmanned systems, military applications of artificial intelligence, and biotechnology developments, along with other emerging technologies, as new frontiers of military struggle and confrontation that could reshape future deterrence.¹⁰ In particular, the development of unmanned systems and artificial intelligence promise new near-term pathways for coercive leverage, and unmanned systems, particularly, are already starting to be applied in pressure operations targeting U.S. and Taiwanese decision-making. In this regard, these trends and themes will likely remain the subject of active debate with significant implications for the Indo-Pacific region and the global military balance.

ENDNOTES

1. There is also the question of how strategic substitution as a concept relates to other concepts prominent in Chinese writings, such as the notion of “asymmetrically counterbalancing” (非寇聳制衡) powerful adversaries that the Science of Strategy highlights. These information-age and other disruptive capabilities can also be characterized as shashoujian (杀手恝), or “assassin’s mace” capabilities in more popular parlance.

2. Phil Stewart and Patricia Zengerle, “Under fierce Republican attack, U.S. General Milley defends calls with China,” Reuters, September 28, 2021.

3. David C. Logan, “Are They Reading Schelling in Beijing? The Dimensions, Drivers, and Risks of Nuclear-Conventional Entanglement in China.” Journal of Strategic Studies 46, no. 1 (2023): p. 5-55. Henrik Stålhane Hiim, M. Taylor Fravel, and Magnus Langset Trøan. “The Dynamics of an Entangled Security Dilemma: China’s Changing Nuclear Posture.” International Security 47, no. 4 (2023): 
p. 147-187.

4. See the initial reporting from Microsoft threat researchers on this activity, among other public sources: https://www.microsoft.com/en-us/security/blog/2023/05/24/volt-typhoon-targets-us-critical-infrastructure-with-living-off-the-land-techniques/

5. Dustin Volz, “In Secret Meeting, China Acknowledged Role in U.S. Infrastructure Hacks,” Wall Street Journal, April 10, 2025.

6. Demetri Sevastopulo, Joe Leahy, Ryan McMorrow, Kathrin Hille, and Chris Cook, “China builds huge new wartime military command centre in Beijing,” Financial Times, January 30, 2025.

7. For further context, see :U.S. Department of Defense, “Military and Security Developments Involving the People’s Republic of China,” December 2024.

8. Looking at the vernacular of U.S. defense debates, the question can also arise as to whether characterizing China’s approach to force development as an “offset strategy” would be appropriate. See, for instance: Timothy Walton, “Securing the Third Offset Strategy,” Joint Force Quarterly 82 (2016): p. 6-15.

9. See for instance: Justin Ling, “Taiwan Is Rushing to Make Its Own Drones Before It’s Too Late,” Wired, June 23, 2025.

10. See the PLA National Defense University’s 2020 edition of Science of Military Strategy (战略学) for further context.