Verdict

Yes, a global collapse of industrial civilization can still be prevented. It is neither inevitable nor the most probable outcome.

However, preserving something resembling the secure, steadily improving industrial civilization experienced by affluent countries during much of the post-1945 period is much less likely.

The most probable outcome is fractured industrial continuity:

• Advanced industrial systems survive globally.
• Wealthy, cohesive and strategically important regions retain sophisticated energy, medicine, computing, manufacturing and military capabilities.
• Numerous poorer, hotter, conflict-prone or badly governed regions undergo temporary or permanent collapse.
• Public services become less reliable even in wealthy countries.
• Infrastructure, insurance, housing and food become increasingly expensive.
• Governments increasingly ration access through prices, borders, surveillance, emergency powers or explicit allocation.
• Global averages may continue improving while security and living conditions deteriorate for large parts of the population.

That outcome is not business as usual, but it is also not a universal return to pre-industrial conditions.

My probability assessment

There is no accepted scientific model capable of integrating climate, war, finance, governance, AI, ecology, pandemics and resource constraints. The following are judgmental ranges, not expert-consensus probabilities.

Global outcome	By 2050	By 2100, approximately following the 2026 trajectory
Managed continuity: industrial civilization remains broadly functional, while absorbing considerable damage	30% central estimate; plausible range 20–40%	25%; range 15–35%
Fractured continuity: industrial civilization survives, but with widespread regional collapses and significant long-term simplification	60%; range 45–70%	55%; range 40–65%
Global systemic collapse: most regions lose reliable modern energy, logistics, public health and industrial maintenance for decades	10%; range 5–20%	20%; range 10–30%

The probabilities would change sharply after a major nuclear war, an engineered pandemic, repeated simultaneous crop failures or a severe AI accident. Conversely, sustained emissions reductions, successful energy diversification, stabilization of great-power relations and serious investment in resilience could push the global-collapse probability substantially lower.

Avoiding all major degradation is now improbable. Avoiding global collapse remains more likely than not.

---

Why global collapse is not presently the base case

The physical situation is serious but not yet equivalent to an unavoidable global carrying-capacity crash.

The period 2015–2025 comprised the eleven warmest years on record, with 2025 approximately 1.43°C above the pre-industrial average. Current policies point toward roughly 2.8°C of warming by 2100; full implementation of current national pledges would produce approximately 2.3–2.5°C. Those are dangerous outcomes, particularly for food, water, heat exposure, ecosystems and coastlines, but they are not scientifically equivalent to a predetermined worldwide termination of industry. (World Meteorological Organization)

Food losses will be substantial. A 2025 empirical study covering six staple crops and roughly two-thirds of global crop calories estimated an average loss of about 120 dietary kilocalories per person per day for each additional degree of warming. Historically observed adaptation and rising incomes offset only a minority of the projected losses. The probability of simultaneous failures in several major breadbaskets is also increasing for important crops. But climate effects are heterogeneous, and food systems retain possible adaptations through breeding, crop switching, storage, trade, irrigation management, dietary change and reduced diversion of crops into animal feed and fuel. (Nature)

Similarly, dramatic tipping scenarios must not be treated as certainties. Substantial weakening of the Atlantic Meridional Overturning Circulation is a serious risk, but recent modelling does not support treating complete collapse during this century as the central expectation. It remains a low-likelihood, high-impact scenario requiring preparation. (Nature)

Industrial society also possesses buffers unavailable to historical civilizations: global scientific knowledge, rapid communication, diverse energy technologies, modern crop genetics, international trade, monetary institutions, emergency logistics, vaccination, desalination, recycling, automation and the ability to redirect enormous industrial capacity. The global response to COVID revealed serious shortages and inequities, but also produced vaccines within a year and reorganized supply and production at extraordinary speed. Global value chains were disrupted rather than destroyed and are increasingly being rewired rather than simply abandoned. (World Health Organization)

COVID was not a “mild” disturbance. It combined a novel pathogen, border closures, workforce disruption, abrupt demand shifts, medical shortages and emergency shutdowns. The fact that the global industrial system survived it is evidence of both vulnerability and considerable adaptive capacity.

---

Why the collapse risk is nevertheless substantial

The danger is not that any one item in the question independently destroys civilization. The danger is that several stresses become synchronized while the capacity to recover is declining.

A useful approximation is:

Collapse risk rises when the frequency and cost of shocks exceed the speed at which institutions, infrastructure, ecosystems and social trust can be repaired.

A flood, epidemic, crop failure, banking crisis or political scandal may be manageable individually. A sequence of them can consume inventories, savings, insurance capacity, government credit, skilled personnel and public patience. Temporary emergency measures then become permanent deterioration.

The central variable is therefore not merely resources. It is recoverable surplus: the energy, money, legitimacy, skills, spare parts, ecosystems and administrative competence left after meeting immediate needs.

---

Assessment of the principal risk systems

1. Climate, food, water, soil and ecosystems: the largest persistent stress

This is the most certain worsening risk cluster. It will not necessarily produce a single global famine, but it will repeatedly damage harvests, infrastructure, health, labour productivity and government finances.

Approximately 1.7 billion people already live in areas where human-induced land degradation reduces crop yields. Around half the global population experiences severe water scarcity during at least part of the year, while roughly two billion people lack safely managed drinking water. Agriculture is particularly exposed because approximately 20% of cropped land is irrigated but produces around 40% of agricultural output. (FAOHome)

The largest food danger is not gradual average yield loss alone. It is the following sequence:

1. Several exporting regions experience heat, drought, flood or plant disease.
2. Prices rise.
3. Exporting governments restrict exports to protect domestic consumers.
4. Import-dependent states exhaust foreign currency and subsidies.
5. Urban food prices trigger unrest.
6. Governments fall, default or become more repressive.
7. Conflict, migration and transport disruption reduce the following harvest.
8. The original shortage becomes a political and logistical famine.

Climate, biodiversity, food, water and health cannot be treated as separate independent probabilities. They are different expressions of the same damaged ecological system. IPBES’s nexus assessment emphasizes that deterioration in one of these domains commonly intensifies the others. (IPBES Files)

Freshwater is especially dangerous because it is geographically local. Money and global food trade can compensate for local shortages, but water itself is costly to move in bulk. Desalination can support affluent coastal cities, but it cannot economically replace the enormous quantities used by continental agriculture. Aquifer depletion, salinization and damaged watersheds are slow ratchets: recovery may require decades or may be impossible on human timescales. The 2026 UN water assessment describes intensifying scarcity and inequality, while related work now uses the term “water bankruptcy” for basins where persistent consumption exceeds renewable supply. (UN Water)

Biodiversity loss matters less because humanity might miss hypothetical future drugs than because ecosystems perform present-day functions: pollination, pest regulation, soil formation, fisheries production, water purification, flood moderation and disease regulation. Losses are often nonlinear and difficult to reverse.

Most probable outcome: progressively higher food-price volatility and regional food emergencies, not an immediate global agricultural shutdown. A global food catastrophe becomes plausible when climate failures coincide with major war, fertilizer disruption and export restrictions.

---

2. Fossil energy, fertilizers, minerals and phosphorus: transition failure is more dangerous than imminent physical exhaustion

Physical exhaustion of all fossil fuels is not a credible near-term collapse mechanism. The world has enough coal, oil and gas to create far more warming than is compatible with a stable climate.

The more realistic risks are:

• declining quality and rising extraction cost in some oil provinces;
• reliance on a small number of exporters and maritime chokepoints;
• sanctions, war or domestic prioritization by producing countries;
• underinvestment in fossil supply before replacement systems are ready;
• slow construction of grids, storage, generation and transmission;
• dependence on concentrated mineral refining and specialized industrial equipment.

Fossil fuels still supply roughly 86% of primary energy, even as clean-energy investment has become approximately twice fossil investment. Under the IEA’s stated-policies scenario, oil demand peaks around 2030, meaning that demand transformation may become more important than geological exhaustion. (Energy Institute)

The transition creates its own concentration risks. The IEA reports that China is the leading refiner for nineteen of twenty strategic minerals it monitors, with an average share around 70%. Mines, refineries, transformers, high-voltage equipment, semiconductors and skilled workforces cannot be produced instantly. (IEA)

Nitrogen fertilizer is primarily produced through ammonia synthesis using hydrogen commonly derived from natural gas or coal. Its dependency is therefore not on “petrochemicals” in the narrow sense but on industrial energy, chemical plants and international transport. It is technically possible to produce ammonia using electrolytic hydrogen, but rapidly replacing the existing global system would require large amounts of electricity, capital and equipment.

Phosphorus is biologically irreplaceable, but claims of an imminent physical “phosphorus cliff” are not well supported. Mineral reserves are economic quantities that change with price, technology and exploration. The nearer risks are geographical concentration, cost, pollution, inefficient use and failure to recover phosphorus from manure, sewage and food waste. Phosphate rock is treated as a strategically critical mineral precisely because disruption and concentration matter before literal exhaustion. (USGS)

Most probable outcome: episodic energy and fertilizer shocks, accelerated electrification and strategic industrial policy. Collapse becomes possible if fossil supplies contract politically or economically before replacement energy, grids and industrial heat are ready.

---

3. Governance, corruption, institutional trust and political legitimacy: the decisive multiplier

Environmental hazards become civilizational hazards when governments cannot make credible decisions, collect taxes, maintain infrastructure, suppress corruption or distribute burdens fairly.

Democratic deterioration is real. V-Dem reported that almost a quarter of countries were undergoing autocratization in 2025, while Freedom House recorded a twentieth consecutive year of declining global freedom. But the latter also found substantial institutional persistence: more than 85% of countries rated “Free” in 2005 remained so two decades later. The evidence therefore supports serious deterioration, not universal democratic disintegration. (V-Dem)

Trust is also uneven rather than simply disappearing. OECD data show low average confidence in national governments and widening trust gaps between government supporters and opponents. Perceived unfairness, discrimination and low socioeconomic status are strongly associated with distrust. Only about 42% of respondents across OECD countries were confident that their country would substantially reduce emissions over the following decade, despite around 70% considering climate action a government priority. (OECD)

This is important because citizens generally do not accept permanent sacrifice merely after being told that it is scientifically necessary. They are especially unwilling to accept it when:

• elites are visibly exempt;
• corruption remains unpunished;
• policies appear ineffective;
• costs are immediate but benefits are distant;
• rival groups are perceived as receiving preferential treatment;
• governments repeatedly reverse policy.

The absolute assertion that citizens have “never accepted less” is historically incorrect. Populations have accepted wartime rationing, taxation, military mobilization, reduced energy use, smoking prohibitions and other restrictions. But durable consent normally requires a visible threat, perceived reciprocity, competent administration and a credible end or compensating benefit.

Democracies therefore face a difficult but not impossible problem. The least realistic strategy is to ask the majority to vote explicitly for indefinite material decline while affluent minorities retain conspicuous consumption. More feasible approaches substitute investment, standards, public infrastructure and progressive pricing for moral appeals to austerity.

Authoritarian regimes are not a reliable solution. They can build infrastructure quickly and suppress opposition, but they also suppress bad news, protect political insiders, make uncorrected policy errors and concentrate succession risks. Democracy is slow; autocracy is brittle.

Religious fundamentalism, ethnonationalism, far-right extremism and revolutionary ideologies are dangerous through similar mechanisms: they convert empirical questions into identity tests, subordinate institutions to loyalists, constrain women and minorities, damage education and create enemies against whom failure can be blamed. Religion itself is not the independent variable; institutional capture and zero-sum identity politics are.

The United States provides a current example of epistemic polarization affecting policy. In 2025–2026, the federal government removed routine COVID-vaccine recommendations for healthy children and pregnant women and attempted or implemented broader reductions in blanket childhood-vaccine recommendations. Parts of the overhaul were challenged and blocked in court. By July 2026, measles cases had reached a 35-year high amid declining vaccination coverage. This is serious institutional damage, but the litigation, state-level resistance and continued operation of medical institutions also illustrate residual checks and resilience. (Reuters)

Most probable outcome: more unstable democracies, more hybrid and authoritarian systems, and greater politicization of technical agencies. Complete state failure remains concentrated in vulnerable states rather than becoming universal.

---

4. Deepfakes, AI-generated content and the loss of shared reality

The concern is valid, but the principal risk is often misunderstood.

Deepfakes need not convince most people to cause damage. Their more powerful effects are:

• making fraud, impersonation and harassment cheap;
• overwhelming verification systems with volume;
• allowing genuine evidence to be dismissed as fabricated—the “liar’s dividend”;
• increasing the cost and time needed to authenticate evidence;
• enabling personalized propaganda and automated influence operations;
• creating election disputes before verification can occur.

Reality does not become literally unknowable. Courts, cryptographic signatures, authenticated sensors, original files, corroborating witnesses, transaction records and chains of custody continue to exist. What changes is that verified reality becomes more expensive, while falsehood becomes inexpensive.

The 2026 International AI Safety Report found documented AI-enabled fraud, scams and manipulation, but stated that real-world political manipulation was not yet widespread. It also found that current systems remain unreliable, that AI agents increase the difficulty of human intervention, and that existing safeguards remain incomplete. NIST is developing provenance, watermarking and forensic approaches, but none is universally reliable. (International AI Safety Report)

Conventional propaganda, selective reporting, partisan elites and social identity remain at least as important as technically sophisticated deepfakes. People frequently believe false claims because trusted group leaders endorse them, not because the media artifact is technically convincing.

Most probable outcome: authentication systems become routine, anonymous media loses evidential value, identity fraud rises and political groups increasingly inhabit different information environments. This seriously damages democratic coordination but is unlikely to collapse industrial civilization by itself.

---

5. Finance, debt, insurance and high-frequency trading

The claim that modern finance “does not tolerate strong disturbances” is overstated. It survived the 2008 crisis and COVID through central-bank liquidity, deposit guarantees, fiscal transfers, regulatory suspension and enormous public balance sheets.

The limitation is that financial policy can create liquidity but cannot manufacture food, electricity, transformers, antibiotics or skilled technicians. It can bridge a temporary physical shortage. It cannot permanently replace missing productive capacity.

The IMF estimated global public debt at just under 94% of world GDP in 2025 and projects it to approach 100% by 2029. This reduces fiscal freedom to respond repeatedly, particularly when rescue spending competes with aging populations, defence, climate adaptation and infrastructure. Leverage outside conventional banks can also produce forced asset sales and liquidity spirals. (IMF)

Insurance is a major underappreciated transmission mechanism. Swiss Re estimated the 2025 natural-catastrophe protection gap at $424 billion. As risks become repeatedly uninsurable:

1. premiums rise or coverage disappears;
2. mortgages become harder to obtain;
3. property values and local tax revenue fall;
4. municipalities defer infrastructure;
5. services deteriorate;
6. residents with money leave;
7. the remaining tax burden rises.

This can produce collapse-like conditions without any dramatic national event. (Swiss Re)

High-frequency trading can worsen flash crashes and cause market liquidity to disappear when it is most needed. It remains a secondary amplifier, not a major civilizational driver. Circuit breakers, clearing rules, trading halts and central banks can contain most purely financial episodes. High-frequency trading becomes dangerous principally when it accelerates a crisis rooted in real insolvency, war or physical scarcity. (Bank for International Settlements)

Most probable outcome: repeated public rescues, financial repression, inflationary episodes, higher taxation and gradual socialization of climate losses. The wealthier core of the financial system is protected, while households, peripheral states and local governments bear much of the adjustment.

---

6. Global supply chains, semiconductors and complex infrastructure

Globalization is simultaneously a source of fragility and a source of resilience.

It creates chokepoints, long production chains and dependencies on highly specialized plants. But it also allows a failed harvest, factory or power system to be replaced by production elsewhere. Complete national self-sufficiency can be less resilient than diversified trade.

The correct resilience model is not autarky. It is:

• multiple suppliers in different geopolitical regions;
• strategic stocks for slow-to-replace items;
• standardized components;
• repairability;
• spare industrial capacity;
• domestic or allied production of essentials;
• emergency transport agreements.

Global value chains remain the backbone of production and are being reorganized rather than eliminated. Semiconductor fabrication, mineral refining, cloud infrastructure, pharmaceuticals, industrial control systems and electrical transformers remain particularly concentrated. Governments are now spending heavily to relocate some semiconductor capacity, itself evidence that the vulnerability is recognized—but new production ecosystems require years, skilled labour and entire networks of suppliers. (World Trade Organization)

Aging infrastructure is a serious ratchet. It is not accurate to say that all global concrete infrastructure is simultaneously reaching the end of its life; ages, standards and maintenance vary enormously. The systemic problem is that maintenance backlogs coincide with climate conditions exceeding original design assumptions. At the same time, governments must finance grid expansion, coastal protection, water systems, digital infrastructure, defence and social care.

Complex systems do not fail merely because they are complex. Complexity also supplies monitoring, specialization, substitution and rapid repair. They fail when critical components lack redundancy and the repair system itself depends on the failed component.

Examples include:

• electrically pumped water without backup power;
• digital logistics with no manual operating mode;
• hospitals dependent on networked records and just-in-time pharmaceuticals;
• grids dependent on transformers with multi-year replacement times;
• communications dependent on a few cloud, cable or satellite systems;
• cities dependent on continuous fuel and food deliveries.

Most probable outcome: greater regionalization and stockholding, higher costs and reduced efficiency. Complete deglobalization would itself lower living standards and make many technologies harder to maintain.

---

7. Cyberattacks and dependence on information technology

Cyber risk is rising because attackers can affect physical systems through industrial control networks, suppliers, remote access and compromised software updates. Water utilities, hospitals, transport and power systems are attractive targets, and many operate legacy equipment that cannot be patched easily. ENISA continues to document a large and evolving threat landscape, while CISA treats water and wastewater systems as vulnerable to cyber and physical attack. (ENISA)

AI will strengthen both attackers and defenders. The 2026 AI Safety Report found that AI systems can discover vulnerabilities and produce malicious code, but it remains uncertain which side will benefit more. (International AI Safety Report)

A cyberattack alone is unlikely to cause permanent continental collapse. Sustained collapse would require several additional failures:

• simultaneous attacks across sectors;
• physical destruction rather than temporary encryption;
• unusable backups;
• lack of manual operation;
• compromised communications;
• inability to manufacture replacement equipment;
• continuing military or political disruption preventing repair.

The realistic objective is not perfect cybersecurity. It is segmentation, offline backups, manual operation, spare parts, trained personnel and rapid restoration.

Most probable outcome: frequent expensive disruptions and occasional regional emergencies. In a major war, coordinated cyber and physical attacks could become a central collapse mechanism.

---

8. Pandemics, climate-sensitive disease and antimicrobial resistance

“Climate-induced pandemics” is too direct a description. Climate change modifies vector ranges, water quality, seasonal transmission, wildlife movement and human exposure. Pandemic emergence also depends heavily on land-use change, intensive livestock production, wildlife trade, laboratory practice, international mobility and public-health capacity. WHO’s One Health work emphasizes these interacting drivers. (World Health Organization)

Climate change is therefore a risk multiplier rather than a universal pandemic generator.

Antimicrobial resistance is more predictable and chronic. A major global analysis estimated approximately 1.14 million deaths directly attributable to bacterial antimicrobial resistance in 2021 and projected approximately 1.91 million annually by 2050 under its reference scenario. AMR is unlikely to collapse civilization by itself, but it gradually makes surgery, childbirth, cancer treatment, intensive care and routine infection management more dangerous. (PubMed)

A more important omission from the question is engineered biological risk. Advances in biotechnology, automation and AI may lower informational barriers to pathogen modification. The 2026 AI Safety Report found that some frontier systems could provide expert-level biological and chemical information, although material, tacit-knowledge and laboratory barriers remain substantial and uncertain. (International AI Safety Report)

Most probable outcome: more outbreaks, higher health expenditure and continuing AMR mortality. A deliberately engineered pandemic remains lower probability but potentially much more destructive than ordinary climate-mediated disease shifts.

---

9. Migration, inequality, poverty and social unrest

Climate migration is real, but projections are often misinterpreted. The World Bank’s figure of up to 216 million internal climate migrants by 2050 is a high-impact scenario, not a prediction of 216 million international refugees. Most climate-related movement is expected to be internal or regional. At the end of 2025, approximately 82.2 million people were internally displaced, predominantly by conflict, with disasters accounting for a smaller but substantial share. (World Bank)

Migration stresses receiving societies through housing, schools, transport, employment and political reaction. It does not mechanically reduce food availability. Migrants also work, pay taxes, produce food and offset aging workforces. The outcome depends on speed, scale, labour absorption, housing construction, social policy and whether political actors exploit the issue.

The most dangerous migration cascade is:

1. environmental and political failures displace people;
2. neighbouring states close borders;
3. informal camps and smuggling expand;
4. receiving populations perceive loss of control;
5. extremist parties gain;
6. humanitarian and integration systems are cut;
7. conditions worsen and further radicalize both migrants and hosts;
8. international cooperation falls.

Inequality contributes because it weakens common sacrifice. A society can survive considerable poverty when burdens are seen as unavoidable and shared. It becomes unstable when lower groups experience deteriorating conditions while insulated elites accumulate visible gains and influence policy.

Most probable outcome: severe regional displacement, hardened borders, selective migration and recurring political crises. Rich countries will generally retain the physical capacity to absorb more people than their political systems are willing to accept.

---

10. Pollution, toxic legacies and nuclear waste

Air, soil and water pollution already impose an enormous chronic burden. WHO estimates approximately 6.6 million annual deaths from household and ambient air pollution. PFAS and other persistent chemicals accumulate in ecosystems and human bodies, while evidence about endocrine-disrupting mixtures remains incomplete. (World Health Organization)

These are generally health, fertility and productivity drains rather than primary global-collapse mechanisms. They become more serious when public-health systems weaken and contaminated water, industrial sites or waste repositories are no longer managed.

Dumped munitions, abandoned mines, damaged reactors, sunken nuclear vessels and poorly contained hazardous waste are genuine long-term liabilities. Most are geographically bounded. Their global systemic importance is lower than climate, war, food, water or state failure.

The principal nuclear danger is nuclear weapons, not ordinary radioactive waste. Civil nuclear sites require long-term institutional stewardship, but a nuclear war could simultaneously destroy cities, grids, logistics, agricultural trade, state institutions and the climate conditions required for food production.

There is evidence that international chemical regulation can work. UNEP monitoring has found declines in several persistent organic pollutants regulated under the Stockholm Convention, while replacement chemicals create continuing problems. (UNEP - UN Environment Programme)

---

11. Space debris and solar events

A severe geomagnetic storm can damage satellites, disrupt GPS and induce currents in power grids and pipelines. The May 2024 G5 storm caused significant agricultural GPS disruption, with estimated losses exceeding $500 million for US farmers. The most severe scenarios could damage hard-to-replace transformers, particularly where grids are poorly protected. (NOAA Satellite Data)

This is a low-frequency but credible risk. Grid hardening, transformer reserves, operational procedures and geomagnetic monitoring can greatly reduce it.

Kessler syndrome is also real, but its consequences are frequently exaggerated. An uncontrolled collision cascade could render some low-Earth-orbit bands increasingly dangerous and degrade Earth observation, communications and navigation. It would not remove all satellites in every orbit, terrestrial communications, fibre networks or industrial knowledge. ESA concludes that active debris removal will be required to avoid continuing deterioration of the orbital environment. (European Space Agency)

Collapse weight: low as a standalone risk; medium as an amplifier during war or infrastructure failure.

---

12. AI weaponization and loss of control

Immediate AI dangers are more empirically grounded than full runaway scenarios:

• cyber assistance;
• fraud and impersonation;
• biological information;
• autonomous targeting;
• accidental military escalation;
• unreliable software in critical systems;
• labour-market disruption;
• concentration of economic and informational power.

Current systems do not possess the capabilities required for a full loss-of-control scenario, according to the 2026 International AI Safety Report. They are, however, improving in relevant areas such as autonomous operation, software engineering and circumvention of evaluations. The future capability trajectory is highly uncertain, and existing safety arrangements remain largely voluntary. (International AI Safety Report)

It is incorrect to assign runaway AI zero probability merely because current systems cannot do it. It is equally incorrect to treat it as an already imminent certainty.

A particularly dangerous pathway is not a self-aware machine suddenly attacking humanity. It is states and corporations delegating progressively more authority to competitive systems because doing so provides military or economic advantage, until meaningful human supervision becomes too slow or costly.

Collapse weight: medium current risk through misuse and concentration; potentially very high future tail risk, with unusually deep uncertainty.

---

13. Geoengineering

Solar-radiation modification could theoretically reduce global temperature quickly, but it would not reduce atmospheric greenhouse-gas concentrations, ocean acidification or all regional climate effects. Modelled hydrological consequences would be geographically unequal. Sudden termination after long deployment could produce rapid warming, and disputes over responsibility for droughts or floods could create geopolitical conflict.

As of 2026, large-scale deployment remains speculative. UNEP considers it poorly understood, potentially disruptive to climate, biodiversity and the ozone layer, and no substitute for emissions reduction or adaptation. (UNEP - UN Environment Programme)

The largest near-term risk is political: worsening climate conditions may cause one state or coalition to deploy a system with global consequences before governance arrangements exist.

Collapse weight: currently low, but rising if mitigation fails badly.

---

14. Education and human capital

There is no solid basis for saying that education quality is uniformly declining everywhere. Access and outcomes vary considerably. Nevertheless, global education systems are under pressure from conflict, climate disasters, fiscal austerity, teacher shortages and reduced international aid. At least 242 million students experienced schooling disruption from extreme climate events in 2024, while UNESCO expects substantial reductions in international education assistance. (UNICEF)

Educational decline is a slow multiplier:

• fewer technicians and engineers;
• weaker public-health literacy;
• lower capacity to distinguish evidence from propaganda;
• reduced administrative competence;
• lower productivity;
• easier political manipulation.

AI may improve access to tutoring and expertise, but excessive dependence can weaken independent reasoning and make education quality dependent on a handful of private platforms.

Collapse weight: low as a trigger, high as a decades-long erosion of recovery capacity.

---

Important risks omitted or underweighted in the question

Nuclear war and conventional great-power war

This is the most important omission. A large nuclear war can create direct destruction, electromagnetic effects, trade collapse and worldwide agricultural disruption in a short period. Even a major conventional war involving maritime chokepoints, satellites, cyber systems and sanctions could synchronize energy, fertilizer, chip and food shortages.

SIPRI reported in 2026 that states are placing greater reliance on nuclear weapons while arms-control constraints are weakening and modernization continues. (SIPRI)

Engineered pandemics and laboratory accidents

Natural spillover is not the only biological risk. Biotechnology is becoming more capable and geographically distributed. Defensive research and medical manufacturing are improving as well, but governance and verification are weak.

Climate-insurance-property-local-government feedback

This is likely to become one of the earliest visible forms of collapse in wealthy countries: locations remain physically habitable but become financially unmaintainable because insurance, mortgages and public infrastructure cease to function normally.

Ocean warming, acidification, deoxygenation and fisheries

Ocean changes affect protein supplies, coastal economies, storm damage and ecosystems. They are partly independent of terrestrial crop risk and are not corrected by solar-radiation modification.

Crop genetic concentration and livestock or plant disease

Industrial agriculture depends on a limited number of crops, breeds, seed suppliers and standardized practices. A highly successful plant pathogen or livestock epidemic can combine with climate stress and trade restrictions.

Demographic imbalance

Very low fertility and aging strain pensions, healthcare, military recruitment and maintenance workforces in some countries. Other countries have large young populations without adequate employment. Migration can partly match these complementary imbalances, but political barriers are high. Low fertility also reduces long-term ecological pressure, so it is not a purely negative risk. (UN Population Division)

Semiconductor, cloud and communications concentration

A small number of firms and regions supply advanced chips, lithography equipment, cloud computation, operating systems and industrial software. The issue is not only availability but loss of the skills required to reproduce them.

Major volcanic eruptions and asteroid impacts

These are low-probability natural risks. A major volcanic event is more relevant than ordinary cold weather as a possible source of several consecutive poor harvests. Neither materially changes the central forecast, but both justify monitoring and contingency planning.

---

The most important cascading and ratcheting mechanisms

The food–politics–trade cascade

Crop loss causes prices and export restrictions; export restrictions intensify shortages; shortages destabilize governments; instability disrupts the next agricultural cycle.

The disaster–insurance–municipality cascade

Repeated disasters eliminate insurance; property values and tax bases fall; infrastructure maintenance is cut; service failure accelerates departure and fiscal decline.

The distrust cascade

Institutions make mistakes or conceal misconduct; trust falls; citizens refuse valid guidance; policy performs worse; the failure confirms distrust. AI-generated content accelerates this loop but does not create it.

The debt–maintenance cascade

Emergency spending and high interest costs displace preventive maintenance. Deferred maintenance raises the cost of later disasters, requiring more borrowing and further cuts.

The migration–polarization cascade

Displacement stresses housing and services; political entrepreneurs blame outsiders; integration is deliberately weakened; social conflict then makes integration more difficult.

The security-dilemma cascade

States stockpile, subsidize domestic industry and restrict exports to increase security. Other states interpret this as hostility and respond similarly. Global efficiency falls, prices rise and military competition consumes resources needed for adaptation.

The ecological ratchet

Soil loss, groundwater depletion, species extinction, ice-sheet loss and some contamination are not repaired on electoral or normal investment timescales. Each delay permanently narrows future options.

The skills and institutional-memory ratchet

When skilled workers emigrate, retire or leave failing sectors, replacing physical assets is insufficient. The tacit knowledge needed to operate them may take years to recreate.

These mechanisms explain why collapse is more likely to be gradual and discontinuous than a single event: deterioration interrupted by partial recoveries, followed by another shock before the previous damage has been repaired.

---

Stabilizing mechanisms that catastrophist accounts commonly omit

Treating only positive feedbacks is as misleading as ignoring them.

High prices stimulate efficiency, substitution, recycling and new supply. Renewable generation reduces exposure to imported fuels. Trade can replace regional harvest failures. States can create money, ration scarce goods, requisition capacity and suspend normal rules. Scientific knowledge can be copied rather than mined. Demographic growth is slowing. Disasters sometimes create political mandates for reform.

International cooperation has produced major successes when the objective is measurable, verification is possible, substitutes exist and costs can be distributed. The Montreal Protocol has phased out almost all controlled ozone-depleting substances and put the ozone layer on a recovery trajectory. Smallpox was eradicated through cooperation between rival states, including the United States and Soviet Union. (UNEP - UN Environment Programme)

These successes do not prove climate cooperation will work. Climate is harder because nearly every economic activity is involved, benefits are delayed, costs are politically visible and states can free-ride. But they disprove the claim that humans are categorically incapable of long-term international action.

Historical collapse research also does not support simple determinism. Societies exposed to similar environmental or political stresses have experienced widely different outcomes, including severe collapse, reorganization and successful adaptation. Collapse usually has multiple interacting causes, and institutional response matters greatly. (PMC)

---

What can realistically be done

The realistic objective is not to eliminate every risk. It is to prevent synchronization, preserve recovery capacity and keep essential systems operating during shocks.

Measures for which the technology is substantially known

These include:

• rapid expansion of low-carbon electricity, grids, storage and firm generation;
• maintaining sufficient transitional fuel and fertilizer security while alternatives scale;
• strategic reserves of grain, fertilizer, medicines, transformers and critical components;
• diversified suppliers rather than blanket reshoring;
• soil conservation, crop diversification and seed banks;
• aquifer regulation, leak reduction, wastewater reuse and more efficient irrigation;
• heat-resistant buildings and urban cooling;
• microgrids and backup power for water, hospitals and communications;
• manual operating modes and cyber segmentation;
• disease surveillance, vaccine manufacturing and antimicrobial stewardship;
• protection and restoration of watersheds, wetlands and coastal ecosystems;
• anti-corruption enforcement and transparent public procurement;
• progressive compensation for transition costs.

The principal constraints are political distribution, institutional competence and financing—not absence of knowledge.

Measures where the direction is known but implementation remains technically or institutionally incomplete

These include:

• scalable low-carbon fertilizer;
• affordable seasonal energy storage;
• verification and governance of advanced AI;
• synthetic-media provenance;
• engineered-pandemic prevention;
• large-scale carbon removal;
• international migration allocation;
• adaptation of financial and insurance systems;
• governance of solar geoengineering;
• active orbital-debris removal.

Risks for which prevention is much easier than recovery

These include:

• nuclear war;
• widespread aquifer exhaustion;
• major soil loss;
• species extinction;
• uncontrolled orbital debris cascades;
• release of highly capable engineered pathogens;
• deployment of inadequately controlled advanced AI;
• irreversible climate tipping processes.

Once triggered, there may be no reliable remediation. This justifies precaution despite uncertain probability.

---

What is politically plausible

A globally coordinated, egalitarian, planned reduction in material consumption is not politically probable.

A more realistic path is a mixture of:

• national-security-driven energy investment;
• industrial policy and strategic stockpiles;
• clean technologies adopted because they become cheaper or more convenient;
• adaptation after visible disasters;
• stricter building and vehicle standards;
• public investment that reduces household costs;
• partial redistribution to maintain consent;
• regional rather than universal international agreements;
• emergency rationing during actual shortages;
• increasing coercion and surveillance.

Governments will generally act late. They will often protect incumbents and strategic sectors first. Some adaptation will occur only after preventable deaths and asset losses. Poor countries will receive insufficient assistance. Rich regions will protect themselves through borders, subsidies and infrastructure, while externalizing losses.

That is the hard political baseline.

Nevertheless, citizens do not have to accept an equal reduction in welfare for material throughput to fall. Electrification, durable products, smaller vehicles, efficient buildings, public transport and reduced waste can provide similar services with less energy and material input. But claims that the transition requires no sacrifice whatsoever are also false. Some activities—frequent aviation, very large homes and vehicles, high meat consumption, disposable goods and continual expansion of infrastructure—cannot be generalized to the entire world within ecological constraints.

The political requirement is not universal altruism. It is a system in which enough people believe that:

1. the measures work;
2. others are also contributing;
3. elites are not exempt;
4. essential living standards are protected;
5. losses from inaction are visibly greater.

OECD evidence indicates that perceived fairness and effectiveness are central to durable climate-policy support. (OECD)

---

The likely sequence from 2026 onward

Through approximately 2040

Clean electricity and electrification continue growing, but fossil use remains substantial. Climate disasters and heat intensify. AI increases productivity in some sectors while worsening fraud, information overload and labour insecurity. Governments spend more on defence, industrial policy, disaster recovery and border enforcement. Regional food and water crises increase, but the global system remains operational.

Approximately 2040–2070

The adaptation divide becomes more pronounced. Insurance withdrawal, coastal defence, water scarcity and heat make some locations economically unviable. Food-price volatility increases. Governments choose which infrastructure to protect and which communities to abandon or relocate. Trade becomes more bloc-based. Some states lose capacity; others become more technologically sophisticated and coercive.

Late century

The outcome depends heavily on peak warming, great-power relations and whether early-century infrastructure was rebuilt.

A world stabilizing near the lower end of current projections, without major war, remains industrial and highly capable, though environmentally damaged and unequal.

A world above approximately 3°C, with repeated agricultural shocks, weak states, arms races and fragmented trade, could cross into sustained global contraction. Collapse would likely progress from peripheral and import-dependent regions toward more resilient cores rather than occurring everywhere simultaneously.

---

Bottom line

There is still a real and substantial chance to prevent global industrial collapse.

But three propositions should be rejected:

1. “Technology and markets will automatically solve it.” They will solve some constraints and create new ones. Political coordination and public infrastructure remain indispensable.

2. “The number of risks makes collapse mathematically inevitable.” Many listed risks overlap and are not independent. Several are secondary amplifiers rather than primary drivers. Adding their individual probabilities would grossly exaggerate the total.

3. “Because global collapse is not the most likely outcome, civilization is broadly safe.” The modal outcome is likely to include regional collapses, mass displacement, significant mortality, loss of services, authoritarian responses and abandonment of vulnerable communities.

The hardest realistic conclusion is:

Industrial civilization will probably survive the twenty-first century, but it is unlikely to remain universal, steadily improving, politically liberal or equally accessible.

The best-organized states and blocs are likely to preserve advanced industrial complexity. The weakest regions will suffer first and most severely. Some will experience exactly the protracted simplification described in the question. Whether those regional failures spread into a genuinely global collapse depends less on any single resource running out than on whether war, ecological damage, distrust, fiscal exhaustion and infrastructure failure become synchronized faster than societies can repair them.

As of mid-2026, that synchronization is a serious risk, not an accomplished fact.