Are We Already in a Third World War?

By the spring of 2026, the question of whether the world is heading towards a global systemic conflict can no longer be dismissed as alarmist speculation. It has become a legitimate subject of strategic analysis. Across multiple theatres, interstate conflict has intensified, regional wars have begun to overlap, major powers are directly or indirectly engaged in active hostilities, and the institutions designed to contain escalation are failing to do so. The risk is not a repeat of the global conflicts of a 20th century, but the emergence of a new configuration of systemic conflict.

The Middle East illustrates the speed with which local wars can expand into broader confrontations. What began as Hamas's attack on Israel on 7 October 2023 became Israel's war in Gaza, widened to Lebanon, the Red Sea, Iraq and Syria, and has now seen the United States conduct direct military action in the region, with ongoing strikes on Iranian energy infrastructure and retaliatory actions affecting regional states and maritime traffic through the Strait of Hormuz. The ongoing disruption of that waterway, one of the world’s most critical energy chokepoints, demonstrates how regional conflict can have immediate global political, economic and security consequences.

In Europe meanwhile the largest land war since 1945 enters its fifth year with no settlement in sight. Russia's invasion of Ukraine launched in February 2022 continues with sustained attritional combat and no durable settlement in sight. Since its outbreak, the war has repeatedly generated concerns about escalation involving NATO territory, cyber operations, and nuclear signalling. In South Asia, India and Pakistan, two nuclear-armed states, have experienced recurring crises including the exchange of missile strikes in 2025, raising questions about the reliability of mutual assured destruction as a stabilizing mechanism.

These conflicts differ in origin, geography, and political logic. Yet they share several features that are strategically significant: the participation of nuclear powers; the involvement of both superpowers and middle powers, increasing links between military confrontation and economic coercion, pressure on critical trade and energy routes, and weak or contested mechanisms for diplomatic de-escalation and resolution. The cumulative effect is an international environment in which separate crises are more likely to interact than remain contained, and international policy and partnership is beginning the transition from the prevention of war to the containment of war.

Two further conditions exacerbate the current dynamic. Financial markets have continued to rise through active wars and economic disruption, suppressing the alarm mechanisms that would otherwise generate pressure for de-escalation and mispricing systemic risk. And no counterbalancing coalition exists to signal that the window for course correction is closing: Europe is rearming slowly, China is not ready to play a stabilising role, and the Global South lacks the weight to impose costs on revisionist actors. This paper addresses a question hereforto treated cautiously by analysts: are present conditions comparable to those that preceded (and produced) the world wars of the twentieth century, and if so, in what respects? In addressing this question, it also examines a further uncomfortable and perhaps equally contentious issue: whether the United States, increasingly abandoning its role as the guarantor of of the post-1945 order, is on balance increasing or decreasing the systemic risk of World War III. That question matters because the behaviour of the system’s most powerful actor has disproportionate consequences for global stability.

The evidence and its implications are set out in the sections that follow, working through the evidence in four stages: (i) the structural mechanisms leading to the current predicament, traced from 1989 to the present day; (ii) what the roads to 1914 and 1939 actually looked like, what structural guardrails exist today that did not then, and what new risks the present introduces. (iii) an assessment of the major flashpoints, , Russia, Israel, China, India-Pakistan, and the United States, each examined alongside the most plausible escalation pathway through which a third world war could emerge; and finally (iv) a conclusion on whether the present moment resembles 1914, 1939, or something new, and what this implies in terms of potential solutions

How We Got Here; The Structural Mechanisms of Crisis, 1989–2026

The current crisis did not arrive suddenly. It is the product of a thirty-year accumulation of successes, structural failures, strategic miscalculations, and political choices whose consequences are now interacting simultaneously. Understanding the sequence matters, because the argument that today is genuinely dangerous rests not on any single event but on the way multiple pressures have compounded over a period of decades to produce a metacrisis in which the system is no longer capable of managing the shocks it faces. . The timeline below is organised around two dimensions through which that compounding occurred since 1989, which marked the end of the Cold War and the hinge point for the global expansion of the liberal international order.

The first is the gradual erosion of consequences: a succession of violations of the rules-based order was met with insufficient costs, emboldening the next violator and producing geopolitical moral hazard at scale. The second is the hollowing of institutional safeguards; the progressive degradation of the multilateral organizations, forums, treaties, and communication channels through which crises were historically slowed, contained, or resolved. These mechanisms operated in parallel and reinforced each other. Their combined endpoint, a system simultaneously under maximum stress and stripped of its capacity to respond, is the condition of the world in 2026.

Table 1: Seeds of the current crisis

While each event was significant, it is the cumulative effect that is material in creating momentum towards crisis and a more chaotic endpoint.

Each entry is not merely a discrete crisis but a reduction in the system’s capacity to manage the next one. When Russia’s invasion of Crimea in 2014, the first forcible seizure of European territory since 1945, was absorbed through sanctions the violator calculated it could manage, it sent the clear signal that enforcement carries bearable costs to every other potential revisionist actor.¹ Each subsequent act without systemic consequence, the full invasion of Ukraine in 2022, the demolition of Gaza, the escalation of the Gaza campaign across the region, the US-Israel strikes on Iran, further emboldens these actors, the structural consequence of repeated decisions that collectively establish that the order can be renegotiated.²

Similarly the record from 1989 onward is one of compounding hollowing of institutional safeguards, some members withdrawing from the commitments that sustained them, others exploiting their mechanisms to obstruct rather than resolve, the cumulative effect compounding with each passing crisis.³ The Security Council was vetoed into paralysis on Syria from 2011 and was not able to address Israel in Gaza.⁴ The WTO dispute settlement mechanism was rendered inoperable by US blockage of appellate appointments from 2017.⁵ The INF nuclear arm streaty was abandoned in 2019. New START nuclear weapons agreement was suspended by Russia in 2023. Each of these actions foreclosed another avenue through which states had historically slowed escalation, convening, deliberating, and finding diplomatic exits. As of 2026 none of those avenues remain open. Institutional failure is the inevitable end point of both mechanisms. Each violation that goes unenforced further delegitimises the institutional forum through which enforcement might otherwise be organised. Each institutional degradation reduces the friction that would otherwise slow the next violation. Today the system has arrived at what appears to be an endpoint: maximum simultaneous stress, with no functioning forum capable of containing any of it. Thus, the Security Council veto on Hormuz is not a failure of the moment. It is the logical endpoint of a thirty-year trajectory.

The current polycrisis has therefore been assembled gradually, by multiple actors, across multiple decades. That is precisely what makes it so difficult to resolve.

The Roads to 1914 and 1939.

Historical Analogues, Structural Guardrails, and New Risks

The value of historical comparison is not to claim that history repeats itself. It is to identify which structural mechanisms are operating in the present, and how dangerous the combination is. The roads to both previous world wars are instructive, but the comparison has three components that the existing literature tends to conflate: how the present moment resembles 1914 and 1939; what structural guardrails exist today that did not exist then (and how effective they are); and what genuinely new risks the present introduces. All three matter.

1914 – War without Choice

The catastrophe of 1914 emerged from a system so tightly coupled that a regional assassination produced a general European war within six weeks. ⁶The trigger was not exceptional, the preceding decades had seen the assassinations of heads of state across Europe and America without systemic consequence. What was exceptional was the system's response to it. Rigid alliance commitments converted bilateral crises into multilateral ones. Mobilisation timetables rewarded whoever moved first. Communication between the major powers was inadequate at precisely the moments it mattered most.

The mechanism is cascade without design: local triggers activating pre-committed responses, each response creating a new trigger for the next actor, the system accelerating beyond the capacity of any single actor to arrest it. The structural logic of 1914 is that of an over-constrained problem: more binding constraints than variables available to satisfy them. When that condition holds, satisfying all requirements simultaneously is impossible. Each actor optimises for its own subset of constraints, alliance obligations, mobilisation timetables, domestic political survival, and the aggregate of individually rational decisions produces a collective outcome that no single actor chose or could have prevented. As a result, no major power sought the war it got.⁷

1939 – War of Choice

1939 was different. A state with a coherent ideological programme of expansion had decided the existing order was illegitimate.⁸ The failure was not structural but political, the democratic powers' repeated choice to accommodate rather than confront, each accommodation lowering the perceived cost of the next demand. The mechanism is graduated appeasement: each concession that averts immediate conflict makes the next confrontation harder to win. By the time Poland was invaded, the capacity for collective resistance that might have worked in 1936 had been spent.⁹

2026 - What the present moment borrows from both

The Middle East escalation sequence has a similar structural logic to the events of 1914. The constraints binding each major actor are individually coherent but collectively irreconcilable, and the system is moving toward a non-functional solution because no actor has the resources or clout required to avoid it. Events began with Hamas’ attack in October 2023 which drew a response of such scale that it successively drew in Lebanon, Yemen, Iraq, Syria, and ultimately the United States as direct belligerents, without any of these actors at any stage choosing the full outcome of the sequence they initiated. Whether that cascade was entirely undesigned is a separate question: there is a reasonable analytical case that Israel calibrated its campaign to create conditions under which American entry became probable, although this interpretation is contested. The same can be argued of Hamas in drawing in its own allies. But even if so, the cascade mechanism operated: a sequence of individual decisions produced a collective outcome that no actor chose. That is structurally analogous to 1914.

The accommodation of Russian revisionism from 2014 onwards on the other hand exhibits a dynamic similar to that of 1939. Each Russian act has been met with calibrated protest rather than resistance sufficient to alter the cost-benefit calculation: Crimea drew sanctions, the full invasion in 2022 drew weapons and isolation but not direct confrontation, and nuclear threats drew rhetorical responses but no doctrinal shift. Each round of international responses, while framed as prudent risk management, in fact lowered the threshold for the next act. The progressive hollowing of the United Nations institutional architecture mirrors the trajectory of the League of Nations in the 1930s,¹⁰ an institution rendered progressively less capable of doing its job by the very behaviour it was built to constrain.¹¹

What the present has that neither period had requires assessment on two dimensions: the guardrails that raise the threshold for escalation, and the new risks that the historical periods did not face.

Nuclear deterrence is the most important structural difference and perhaps the primary guardrail. The existence of assured second-strike capability across multiple actors raises the threshold for direct great-power war to a level not seen in 1914 or 1939.¹² Unlike the opaque military planning of 1914, where mobilisation timetables and deployment plans were state secrets,¹³ today's nuclear powers are explicitly transparent about their second-strike doctrines and the consequences of crossing the threshold, a transparency that is itself a deliberate instrument of deterrence.

Additionally, formal alliance structures with explicit treaty tripwires create clearer red lines than the ambiguous commitments that produced miscalculation in 1914.¹⁴ NATO's Article 5 is an explicit military commitment that has been repeatedly reaffirmed, and backed by integrated command structures.¹⁵ By contrast, Britain's obligation to Belgium under the 1839 Treaty of London, invoked as the casus belli for its entry into the First World War, was a collective guarantee that required a political decision to invoke, and was accordingly dismissed by Germany's Chancellor Bethmann-Hollweg on the eve of invasion as 'a scrap of paper', with disastrous consequences for Germany.¹⁶

Further, direct leader-to-leader communication infrastructure exists today that had no equivalent in either previous period. Infrastructure like the Washington-Moscow hotline established after Cuba, US-China crisis channels, and the broader architecture of diplomatic back-channels of the Cold War were constructed precisely because miscalculation has proven to be existential. But this infrastructure is not universal: negotiations with Iran on the US-Israel war with Iran had to be routed through Pakistan as an intermediary, with envoys shuttling between meetings and ample room for strategic ambiguity on all sides, indicating that the guardrail itself remains limited to a narrow band of great-power-to-great-power confrontation.

Finally, economic interdependence imposes costs on conflict that were absent in both previous periods. Global trade as a percentage of GDP is currently around 60%,¹⁷ approximately twice the levels seen in 1913 and four times that of the late 1930s, with national prosperity now structurally dependent on the uninterrupted flow of goods, capital, and services across borders in a way that neither pre-war period approached. Further, the current extent of financial system integration, dollar-denominated trade, and global supply chains create additional mutual vulnerabilities that are historically unprecedented. However, while the guardrail is real, whether it can hold under stress is a different matter. The same argument - that the depth of economic integration had made great-power war irrational - was widely accepted in 1914,¹⁸ but subsequent events demonstrate that real costs do not automatically translate into effective deterrence, particularly where the costs fall disproportionately on populations rather than on the political and financial elites whose decisions determine whether wars begin and are financed. Accordingly, these costs have not deterred Russia’s invasion pf Ukraine or indeed, US-Israeli strikes on Iran.

Most importantly perhaps, every guardrail described above presupposes that great powers have a shared interest in maintaining it, an assumption that can no longer be taken as given. The US, architect of the post-war order that produced the guardrails, is increasingly acting in ways that erode them: trade policy premised on coercion weaponises the interdependence that once made conflict irrational, while withdrawal from arms control agreements weakens the transparency on which nuclear deterrence depends. America may well be feeling the pressures and motivations that all great powers in history face in the post-expansion phase as domestic issues become intractable and rivals rise.¹⁹

On the other side of the equation the present introduces new risks that have no historical analogue. Multi-domain warfare, cyber operations capable of disabling critical infrastructure, space-based systems on which military and economic functions depend, undersea cables carrying c.95% of global data,²⁰ all create new potential escalation pathways that could cross the threshold into a hot war before either side has made a conscious decision to do so. The Baltic Sea illustrates the dynamic well: since 2022, suspected Russian vessels operating with deliberate ambiguity have severed a succession of undersea cables and pipelines,²¹ damaging infrastructure without having crossed a threshold compelling a formal military response, indicating that a multi-domain/cyber war between superpowers may already be underway.

Further, the speed of the information environment compresses decision time in ways that the July Crisis of 1914 did not face, which took six weeks to produce general war.²² The promise of AI-assisted military systems raises the possibility of that compression accelerating further still, to a pace at which escalation outruns or compromises human authorisation chains entirely. False alarms which once could be double-checked risk turning into irreversible launch decisions, eliminating the sort of deliberative pause that prevented catastrophe in 1983, when a Soviet early warning system incorrectly detected an American missile launch and a single officer overriding it averted a nuclear counter-strike.²³

Finally, for the first time in history non-state actors today possess the capacity to cause damage, whether through precision munitions, drone swarms, or weapons of mass destruction, on a scale that was previously the exclusive preserve of states, implying a proliferation of actors capable of triggering an escalation cascade to a global conflict. By want of example the Houthis, controlling a fragment of Yemeni territory, in 2024 used cheap drones and missiles to cause a 90% decrease in container shipping through the Red Sea within weeks, disrupting an estimated US$1 trillion in goods,²⁴ and triggering a US-led coalition of over 20 countries to launch sustained offensive military operations inside Yemen.²⁵ Additionally, non-state actors in the form of global corporations, particularly big-tech hyperscalers, increasingly also wield world class economic and military-like power, adding further risks and variables to the problem set.

What makes the combination of these new risks distinctively dangerous is not any one of them in isolation but their intersection with the one condition the guardrails were not designed to manage. Nuclear deterrence, alliance tripwires, and economic interdependence were all constructed on the assumption that the dominant power, the state that built and guaranteed the institutional order, would remain invested in its preservation, an assumption that can no longer be taken for granted. The guardrails of the post war order are load-bearing walls in a structure whose foundation has shifted.

Table 2: maps these structural features across the three periods.

The Flashpoints: Actors, Risks, and Escalation Pathways

Five flashpoints currently shape the risk of a third world war. Each shares a feature that is easy to overlook when each is analysed in isolation: the actors driving escalation in each theatre are moving at the same time. That simultaneity is not coincidence. It reflects a shared, but uncoordinated, reading of the strategic environment; a collective perception that the conditions which previously imposed costs on aggressive behaviour have dissolved, and that the window in which long-suppressed ambitions can be pursued is open now.

The primary driver is the behaviour of the US as the dominant power. Historical analysis of power transitions since Thucydides has consistently identified the declining hegemon as the most dangerous actor in an international system: unlike a rising challenger, a hegemon under pressure perceives its windows as closing and acts with an urgency that a secure power would not.²⁶ That dynamic is operating in 2026, with the US reasserting itself outside the framework it built, and in ways that are appear functionally indistinguishable from the revisionist behaviour it had previously sought to deter.²⁷ That shift has directly licensed the behaviour of others. Russia, Israel, and the United States are each operating on the same implicit logic: that power, not rules, now determines outcomes. That norm has now been formally declared in the current US National Security Strategy, which frames "the outsized influence of larger, richer, and stronger nations" as "a timeless truth of international relations",²⁸ which is now being internalised by every government with revisionist ambitions or existential insecurities.²⁹

Two further conditions sustain the current permissive moment. Financial markets have continued to rise through active wars, Hormuz disruption, and the most aggressive tariff regime in a century, suppressing the alarm mechanisms that would otherwise generate pressure for de-escalation, mispricing systemic risk, and raising the probability of a correction that, when it comes, is likely to be sudden and dramatic.

No counterbalancing coalition yet exists to signal that the window is closing: Europe is rearming slowly and in fragmented fashion, China is not ready to play a stabilising role, and the Global South remains non-aligned and without the weight to impose costs. Unchecked unilateral power eventually produces a coalition to reign it in.The question is what will be made irreversible until that time.

A numer of flashpoints exist in the world that can take the world on a course towards uncontrolled escalation. The key flashpoints are assessed below not by the intensity of current violence but by their capacity to produce uncontrolled escalation drawing in multiple major powers. For each, the actor analysis and its primary escalation pathway are presented together.

Table 3: Flashpoints ranked by systemic escalation risk

Russia

Potential escalation pathway - Baltic Conflict: Russia has consistently read American distraction as opportunity.³⁰The potential escalation path would be that with the United States militarily committed in the Middle East, the moment is structurally like others in which Russian probing of Western resolve has intensified. A mix of hybrid operations, cyber, disinformation, energy disruption, destabilise Estonia or Latvia. Russia, emboldened by public US ambiguity about Article 5 and calculating that its actions fall below the threshold of unambiguous armed aggression, attempts a Crimea-style military annexation of foreign territory. NATO then faces a choice with no good options: respond militarily with kinetic confrontation with a nuclear-armed state or absorb the provocation and destroy the collective defence mechanism that has secured Europe for seventy years.

Israel

Potential escalation pathway, regional state collapse: A potential scenario here begins with sustained US-Israeli military pressure and further targeted killings of Iranian leaders producing Iranian internal fragmentation, a contested succession, a Revolutionary Guard coup, or factional breakdown. A weakened or collapsing Iranian state loses coherent control of its remaining missile and drone arsenals. These pass to non-state actors or are used in a final escalatory strike by a faction with no interest in de-escalation, triggering a regional cascade involving Gulf states, Turkey, and potentially Russia and China as suppliers and diplomatic backers of opposing sides, with no managing power and no coherent state actor to negotiate with. The scenario that produced the 2026 crisis recurs, but without the institutional framework that partially contained it the first time.

China

Potential escalation pathway, Taiwan Strait crisis: A future scenario here centres on Chinese declaration of an air defence identification zone over Taiwan, combined with a naval blockade, that creates a confrontation that neither side can back down from without catastrophic loss of face. From there a Taiwanese move toward formal independence, a red-line US arms transfer, or a domestic Beijing crisis could trigger the sequence, but the timing matters as much as the trigger. The US, already at war in the Middle East and potentially managing a NATO crisis in Europe, cannot credibly prosecute a third simultaneous conflict at full capacity. China has assessed this window. The question is whether it moves before the US recovers its strategic capacity or concludes the moment has passed. A blockade scenario simultaneously threatens the semiconductor supply chains on which the entire global digital economy depends, triggers US treaty obligations, and forces every major power to choose sides in a confrontation that the current institutional architecture has no mechanism to contain.

India and Pakistan

Potential escalation pathway, crisis with compressed decision time: A potential scenario here begins with a terrorist attack attributed to Pakistani state-linked actors occurs in a political environment where Indian domestic pressure for a decisive response is higher and Pakistani civilian political control is weaker than in 2025. The calibrated-strikes-then-ceasefire playbook fails: one side miscalculates the other's red lines, or a faster-moving exchange overtakes third-party intervention before it can operate. Nuclear use becomes a live option before the crisis management architecture can function. The scenario becomes systemic only under one specific condition: Chinese entry, either to support Pakistan directly or to exploit Indian military preoccupation on the western front to press Himalayan border claims simultaneously. A three-actor confrontation draws in US support for India and converts a South Asian crisis into a great-power conflict through the back door without any actor having chosen that outcome. The alternative is that the US is not drawn in, and India is left with the option of using its nuclear capability against both Pakistan and China or suffering a humiliation along the lines of the 1962 Sino-Indian war.³¹

The United States

Primary escalation pathway, Iran conflict escalation: The current Iran campaign has no defined military endpoint, no congressional authorisation, and no exit strategy. The escalation does not require a deliberate decision for wider war, it requires only that one of several observable triggers is crossed. A mass-casualty Iranian strike on Israel. An attack on US naval forces that sinks a ship. An Iranian move toward a nuclear device, no matter how crude, that Israel and the US conclude must be destroyed. Any of these can produces a US escalatory response. Russian and Chinese vetos at the Security Council prevent any potential UN-led de-escalation efforts. . The economic consequences of a prolonged Hormuz closure, 20% of global oil, the arterial route for Asian energy imports, produce political crises across Asia and Europe that constrain governments' capacity to de-escalate even when they choose to, leading to being drawn in to defend their ships rather than join the war. The crisis expands not because any actor chooses world war but because no actor retains the capacity or credibility to stop it.

On the US scenario, with the US increasingly abandoning its role as the guarantor of the global security order, its current strategic posture increasingly resembles that of a revisionist power, rejecting existing international rules, unilaterally asserting territorial claims, economically coercing partners, and withdrawing from multilateral institutions. Whether that revisionist label can bedirectly applied to the US is both debatable and for the purposes of this paper irrelevant. However, America’s status as a flashpoint doesn’t require revisionism in a strict ideological sense. It requires only that its actions are net-destabilising, increasing rather than reducing the probability of cascading escalation. This condition is clearly met today.

The five flashpoint risks are not discrete. Each of the actors involved is a nuclear power, and each is simultaneously an ally, a rival, or a strategic opponent of the others. The escalation pathways described above do not run in parallel, they intersect, trigger each other, and combine as they unfold. Two combination scenarios illustrate the mechanism:

Scenario 1: The Three-Front Trap. Russia moves against the Baltic states while the US is militarily committed in the Middle East. Washington faces a genuine capability choice: sustain its Iran campaign or respond to the NATO crisis with the credibility the alliance requires. It cannot do both at full capacity simultaneously. Europe, still rearming and without the integrated command capacity to substitute for US leadership, cannot fill the gap, and its standalone response to the Baltic crisis is slower, more fragmented, and less credible than Article 5 requires. China observes that the US is now committed across two theatres and concludes the time for action on Taiwan is now. A Chinese blockade of Taiwan does not require coordination with Russia, only a joint reading of the strategic landscape that is visible to everyone. The US now faces three simultaneous theatres, and its deterrence capability in each degrades as its commitment to the others becomes visible. Each adversary independently calculates that the moment is more favourable than it will be when the US recovers the capacity for strategic initiative and escalates accordingly.

Scenario 2: Normalisation of the Nuclear Threshold. India and Pakistan exchange a second more violent and prolonged round of missile strikes. The exchange does not go nuclear, but it comes closer. The US, absorbed in the Middle East, is late and ineffective in its intervention, and China’s engagement is interpreted by India as support for Pakistan in the face of ongoing Sino-Indian border disputes and refused accordingly. While the crisis is eventually contained, the taboo against missile strikes between nuclear-armed states has been normalised as negotiable, and the psychology of nuclear deterrence has been materially eroded. The next crisis, wherever it occurs, begins in a world where nuclear use is less unthinkable than it was before.

The common thread in these scenarios is that no single actor chooses the outcome. Each makes a rational decision given the conditions it faces. The conditions are the product of the compounding mechanisms this paper traces, the erosion of consequences, the hollowing of institutional shock-absorbers, the withdrawal of the system's guarantor.

Conclusions and Implications

The present is not 1914. The cascade mechanism of that crisis, rigid alliance blocs, mobilisation logic that removed decision time, no nuclear constraint, does not precisely describe the world today. It is not 1939 either. There are no major ideological actors pursuing global system replacement. The United States, even in its most nationalist configuration, does not qualify, its behaviour is transactional and self-interest driven, highly damaging to allies and enemies, rather than ideological in the sense that produced the catastrophes of the twentieth century. This moment of danger is distinct from previous world wars, but in many ways harder manage than either. It combines the cascade dynamics of 1914 with the institutional failures of the 1930s, adds nuclear weapons across multiple actors, and does all of this in a system where the dominant power is not managing the order but actively disrupting it. That last feature has no historical precedent and is the most dangerous dynamic of the times.

What makes the current moment especially dangerous is not the intensity of any single crisis. It is their simultaneity, and the absence of any mechanism capable of containing them. In 1962, the Cuban Missile Crisis brought the world closer to World War III than any moment before or since, but it was one crisis, involving two leaders, and a functioning back-channel of communications that allowed the coordination required for both sides to step back from the brink. Today, risk is risk distributed across five simultaneous flashpoints, each with nuclear capabilities, and no actor or institution commands the standing or credibility to perform that moderating function.

The risk of the world inadvertently sliding into a global systemic conflict is therefore currently higher than at any point since 1945. In the absence of deliberate interventions, it will rise further.

In the near term however, y the primary brake on escalation remains the restraint that actors have so far chosen to exercise: Russia has threatened but not yet used a nuclear weapon in Ukraine. China has threatened but not yet moved on Taiwan. India-Pakistan have fired on each other but contained the conflict. Each threshold that has not been crossed remains a decision that can still be made. The question is whether the conditions that have sustained those decisions will hold, and whether deliberate action can reinforce them.

Three levels of intervention could materially reduce the risk this paper has assessed, in descending order of immediate leverage. The most consequential near-term agency lies with the US. A decision to re-engage the institutional frameworks it has been dismantling would do more to reduce cascade risk than any other single intervention available to any actor. The probability of such a course correction is, on current evidence, low.

Second, as a result the security architectures built on the assumption of reliable US institutional leadership, and a functioning multilateral backstop require urgent reassessment. For allies, this is no longer a contingency to plan for, it is the condition they are already operating in. European defence autonomy, Indo-Pacific burden-sharing, and bilateral crisis communication channels between allies are the functional replacements for US security guarantees that can no longer be relied upon. The window for building that architecture before it is needed in a crisis is closing.

Third, the institutional architecture designed to manage moments like this one cannot be reconstructed on the timescale the current risks demand. The most realistic near-term alternative is a coalition of the willing, a group of states with sufficient weight and relationships to convene, back-channel, and broker off-ramps in crises. While such a coalition would be informal and imperfect, it would be more than exists today. The states with the standing and incentive to constitute such a coalition, the EU, India, the Gulf states, Japan, South Korea, and potentially China on specific issues, have not yet acted with the collective urgency the moment requires.

It is time for serious geopolicy thinking and urgency on this topic. While the trajectory to a global systemic conflict is not fixed, the margin for error is narrower than at any point in the post war era, and it is narrowing further.

This article was first published by the Institute for Strategic Intelligence and Intervention in May 2026 and is reprinted here with their permission with minor edits made for style.

The Leader: Endnotes

1. Fazal, T. M. (2025, March 21). Conquest is back: A peace deal in Ukraine could further normalize what was once taboo. Foreign Affairs, 104(2). https://www.foreignaffairs.com/russia/conquest-back/

2. Hathaway, O. A., & Shapiro, S. J. (2025, June 24). Might unmakes right: The catastrophic collapse of norms against the use of force. Foreign Affairs, 104(4). https://www.foreignaffairs.com/united-states/might-unmakes-right-hathaway-shapiro/

3. Ikenberry, G. J. (2020). A world safe for democracy: Liberal internationalism and the crises of global order. Yale University Press.

4. Security Council Report. (n.d.). The veto: UN Security Council working methods. https://www.securitycouncilreport.org/un-security-council-working-methods/the-veto.php

5. Hopewell, K. (2025). Unravelling of the trade legal order: Enforcement, defection and the crisis of the WTO dispute settlement system. International Affairs, 101(3), 1103–1124. https://academic.oup.com/ia/article/101/3/1103/8100243

6. Clark, C. (2012). The sleepwalkers: How Europe went to war in 1914. Allen Lane.

7. Keegan, J. (1998). The first world war. Hutchinson.

8. Kershaw, I. (2000). Hitler 1936–1945: Nemesis. Allen Lane.

9. Overy, R. J. (2016). The origins of the Second World War (5th ed.). Routledge.

10. Patrick, S. (Ed.). (2023). UN Security Council reform: What the world thinks. Carnegie Endowment for International Peace. https://carnegieendowment.org/research/2023/06/un-security-council-reform-what-the-world-thinks

11. Carr, E. H. (1946). The twenty years' crisis, 1919–1939: An introduction to the study of international relations (2nd ed.). Macmillan.

12. Jervis, R. (1989). The meaning of the nuclear revolution: Statecraft and the prospect of Armageddon. Cornell University Press.

13. Tuchman, B. W. (1962). The guns of August. Macmillan.

14. Snyder, G. H. (1997). Alliance politics. Cornell University Press.

15. North Atlantic Treaty Organization. (1949, April 4). The North Atlantic Treaty. NATO. https://www.nato.int/en/about-us/official-texts-and-resources/official-texts/1949/04/04/the-north-atlantic-treaty

16. Clark, C. (2013). The sleepwalkers: How Europe went to war in 1914. Harper.

17. Macrotrends. (n.d.). World trade to GDP ratio. Macrotrends. Retrieved April 28, 2026, from https://www.macrotrends.net/global-metrics/countries/wld/world/trade-gdp-ratio

18. Angell, N. (1910). The great illusion: A study of the relation of military power to national advantage. Heinemann.

19. Patel, K., Hansmeyer, C., Desai, N., & Ajit, A. (2025). American hegemony at a critical juncture: Lessons from history’s great powers. Frontiers in Political Science,7. https://doi.org/10.3389/fpos.2025.1511913 frontiersin.org+1

20. Bischoff, P. (2022). Undersea telecommunication cables: Technology overview and issues for Congress (CRS Report No. R47237). Congressional Research Service. https://www.congress.gov/crs-product/R47237

21. European Council on Foreign Relations. (2025, July 9). Shallow seas and "shadow fleets": Europe's undersea infrastructure is dangerously vulnerable. https://ecfr.eu/article/shallow-seas-and-shadow-fleets-europes-undersea-infrastructure-is-dangerously-vulnerable/

22. Clark, C. (2012). The sleepwalkers: How Europe went to war in 1914. Allen Lane.

23. Schlosser, E. (2013). Command and control: Nuclear weapons, the Damascus accident, and the illusion of safety. Penguin Press.

24. ACLED. (2024). Six Houthi drone warfare strategies: How innovation is shifting the regional balance of power. Armed Conflict Location and Event Data Project. https://acleddata.com/report/six-houthi-drone-warfare-strategies-how-innovation-shifting-regional-balance-power

25. Wilson Center. (2024). Timeline: Houthi attacks. https://www.wilsoncenter.org/article/timeline-houthi-attacks

26. Allison, G. (2017). Destined for war: Can America and China escape Thucydides's trap? Houghton Mifflin Harcourt.

27. Vinjamuri, L., et al. (2025). Competing visions of international order: Responses to US power in a fracturing world. Chatham House. https://www.chathamhouse.org/2025/03/competing-visions-international-order

28. The White House. (2025, December 4). National security strategy of the United States of America. https://www.whitehouse.gov/wp-content/uploads/2025/12/2025-National-Security-Strategy.pdf

29. Institute for Strategic and International Investigations. (2025, December). America the merchant power: The National Security Strategy 2025. ISII. https://www.isii.global/files/pivotal-thinking/america-merchant-power-national-security-strategy-dec-2025.pdf

30. Trenin, D. (2019, February 20). Russia's global ambitions in perspective. Carnegie Endowment for International Peace https://carnegieendowment.org/research/2019/02/russias-global-ambitions-in-perspective

31. Hoffmann, S. A. (1990). India and the China crisis. University of California Press.