A Cold Assessment

One caveat before the substance: my reliable knowledge ends before mid-2026. I'm extrapolating trend lines that have been stable for years and decades, which is legitimate for structural analysis — the geological, demographic, thermodynamic and institutional curves discussed below do not meaningfully change in 18 months — but any specific events of 2025–26 are outside my view.

Now the answer.

Bottom line first

Yes, a chance still exists. It is not the probable outcome, and it shrinks every year.

Three statements that are simultaneously true, and must be held together:

1. Preventing all degradation is already impossible. Committed warming, committed biodiversity loss, committed aquifer depletion, committed demographic inversion, accumulated debt, eroded trust — these are done deals. Whatever happens now, 2100 will be poorer in ecological terms and probably in material security terms than the world of 2015. Anyone promising otherwise is selling something.

2. Preventing full civilizational cascade is still possible — nothing in physics or thermodynamics mandates it. The knowledge and the resources to execute an energy transition, stabilize food systems, and harden critical infrastructure exist today. What is missing is coordination capacity, and that is itself a depleting resource.

3. The modal outcome — the single most likely one — is neither salvation nor apocalypse, but a protracted, uneven, disorderly grinding-down: regional collapses at the periphery (already underway), stagnation and authoritarian drift in the core, periodic compound crises met with decreasingly effective improvisation, and the depth of the eventual floor determined by whether great-power war and climate tipping cascades are avoided or not.

My rough probabilities, defined against your definition of collapse (multi-decade disorderly loss of complexity, safety, food, energy, services):

These are structured judgments, not measurements. The reasoning behind them follows. Note also that the risk is back-loaded: system inertia protects the 2020s and condemns the 2040s–2060s, which is when the climate, shale-decline, demographic, water and AMR curves cross.

The core mechanism: this is not thirty risks, it is one risk with thirty faces

The most common analytical failure — in both doom and optimism literature — is treating your list as a portfolio of independent problems, each with its own probability and its own fix. That is wrong in a specific, describable way.

Every complex system survives shocks through buffers: grain reserves, fiscal space, spare hospital capacity, institutional trust, redundant suppliers, slack in the grid, goodwill between nations. For forty years, the global system has been optimized — deliberately, by markets and by ideology — to remove buffers in the name of efficiency. Just-in-time logistics, just-enough stockpiles, leveraged balance sheets, single-source chokepoints, hollowed-out state capacity, monetized attention. An optimized system converts every shock into a propagating wave instead of absorbing it.

So the true equation is:

Damage = shock magnitude × system coupling × (1 / buffers)

All three terms are moving the wrong direction simultaneously: climate raises shock frequency and magnitude; interconnection raises coupling; efficiency, debt and distrust erode buffers. This is why the individual probability of each risk understates the danger — the risks are synchronized by shared drivers (climate drives food prices, migration, conflict and disease at the same time), and each crisis response consumes the capacity to respond to the next one (debt issued, emergency powers normalized, trust spent). This is the ratchet you asked about. Crises are not independent draws; they are drawn from a deck that gets worse each time you play.

A second structural point: procrastination is not neutral — it reclassifies problems. In 2000, climate was a "we know what to do" problem. A quarter-century of delay has converted part of it into a "nothing to be done" problem (committed sea-level rise, committed ice loss, committed species extinction) and raised the cost of the remainder. The same migration is now occurring with antibiotics (from solvable stewardship to post-antibiotic medicine), with infrastructure (maintenance debt compounding into replacement cost), and with aquifers (regulation possible now; recharging a collapsed aquifer takes centuries). The category "problems we know how to solve" is a wasting asset.

Triage of your list

Rather than re-describing items you already understand, here is the classification that matters:

Class A — Locked in; only adaptation remains. Committed warming to ~1.5–2°C; coral reef loss; a large fraction of biodiversity loss; glacial water-tower depletion for ~2 billion people by mid-century; soil already eroded; pollutants already dispersed (PFAS, endocrine disruptors, the dumped munitions and reactors you mention — these are local toxic legacies, not civilization-enders, but they are permanent claims on future maintenance budgets); demographic inversion (the working-age populations of China, Europe, East Asia are already born or already absent — this is arithmetic, not forecast).

Class B — Solutions fully known; blocked by political economy. Climate mitigation, grid hardening against solar storms (a Carrington-class event is a few percent per decade and the engineering fix is known and unimplemented), pandemic preparedness, antibiotic economics (the subscription model works — Britain piloted it — and remains unscaled), phosphorus recycling, groundwater regulation, soil conservation, bridge and dam renewal, nuclear site decommissioning, food stockpiling, debt reduction. Every one of these is a solved engineering or policy problem. Every one is un- or under-executed because the costs are present and concentrated and the benefits are future and diffuse — the precise configuration democratic electoral cycles handle worst.

Class C — No known solution. Restoring a shared factual reality under conditions of infinite, free, synthetic content (no one has solved this; provenance standards cannot fix the liar's dividend — the mere existence of deepfakes lets any authentic evidence be dismissed); reversing established affective polarization; restoring fertility where it has collapsed (no country has durably succeeded at any price); governing geoengineering; dissolving security dilemmas without trust; running a debt-and-growth-based financial system without growth; aligning superintelligent systems (genuine expert disagreement spanning orders of magnitude — honest treatment means assigning it a small but real and irreversible weight, while noting it is currently a distraction from the boring certainties killing people now, like aquifers).

Class D — Wildcards. AMOC collapse (plausible 10–40% this century on current forcing; would reorganize Northern Hemisphere agriculture and tropical monsoons — one of the few single events that could break global food trade by itself); Taiwan contingency (the global economy's most dangerous single node — ~90% of advanced logic chips; a blockade is a depression-level shock in weeks); Carrington-class solar storm hitting an unhardened grid during another crisis; engineered pandemic (democratized biotech plus AI-assisted design is lowering the barrier every year).

The crucial observation about this classification: items are continuously migrating from B to A through delay, and the system's remaining action-capacity is concentrated in B — which is precisely where nothing is being done. That is the whole diagnosis in one sentence.

What you missed — the risks that belong near the top

Your list is strong but omits several first-order items:

1. Great-power war, including nuclear. This is the single fastest path to collapse and arguably at its highest probability since 1962. The entire post-1945 arms control architecture is dead or dying (INF gone, Open Skies gone, New START expiring with no successor in sight); three simultaneous nuclear-armed confrontations (Russia–NATO, China–Taiwan, India–Pakistan — with Pakistan being a nuclear state facing existential climate-water stress); proliferation pressure on Iran, Saudi Arabia, South Korea, Japan. Recent modeling of a full US–Russia exchange: hundreds of millions dead directly, and nuclear-winter agriculture collapse putting ~5 billion at risk of famine. Even a limited regional nuclear use would shatter the global trading system overnight. Unconditional probability of some nuclear weapon use in anger by 2070: I would put it at 10–20%. That single number should dominate most of your list and almost never does.

2. Demographic inversion. Fertility is below replacement nearly everywhere outside sub-Saharan Africa (South Korea ~0.7; China ~1.0–1.1). This is a double-edged certainty: it relieves resource pressure in the long run, but for the next forty years it means exploding old-age dependency, pension and healthcare claims crowding out adaptation investment, shrinking workforces to maintain the infrastructure, and gerontocratic politics structurally hostile to change. It interacts toxically with migration: the countries that demographically need migrants are politically closing to them.

3. The insurance retreat. Insurers are withdrawing from Florida, California, parts of Australia and Southern Europe. This is the financial system's price signal for climate, and it is a ratchet: when insurance dies, mortgages die; when mortgages die, property values die; then municipal tax bases die; then infrastructure maintenance dies. It is already operating, quietly, and it converts slow climate damage into punctuated fiscal failure.

4. Chokepoint geography. Beyond Taiwan: Hormuz, Malacca, Bab el-Mandeb, the Panama Canal (already drought-limited), and the undersea cable network carrying ~99% of intercontinental data — demonstrably sabotageable, as the Baltic incidents showed. The system's connectivity is concentrated in a handful of narrow places, and state actors have mapped all of them.

5. Processing monopolies. Not mineral depletion — the crust has enough lithium, cobalt, rare earths — but refining concentration: China controls 60–90% of processing for most transition-critical minerals and has already begun export-control weaponization (gallium, germanium, graphite, rare earths). The energy transition runs through a single country's industrial policy. That is a fragility, not a geology, problem — but it is a decade-scale problem to fix.

6. Loss of maintenance knowledge. Complex systems are increasingly operated by aging cadres with no replacement pipeline, while education quality declines. Several near-accidents in recent decades trace to this. A civilization can lose the capacity to run its infrastructure before it loses the infrastructure.

The historical record — read coldly

You asked for pattern-matching against our past. Here it is, without flinching.

When humanity has successfully cooperated against a shared threat, one or more of four conditions held:

• The threat was vivid, fast, and killed elites too (nuclear war after Cuba; smallpox).
• The fix was cheap and required no behavior change (Montreal Protocol — one industry, one substitute chemical; lead removal; acid rain).
• A hegemon underwrote the system (post-1945 institutions under the US umbrella).
• The catastrophe had already happened (the UN, the EU, Bretton Woods — all built on rubble).

None of these four conditions holds today. The threat is slow, statistical, and escapable-by-the-rich in the near term — the worst possible profile for human risk perception. The fixes are expensive and touch every citizen's life. The hegemon is contested and internally paralyzed. And we are pre-rubble.

The deepest and ugliest pattern in the record: large-scale human cooperation almost always follows catastrophe rather than preventing it. The one great counterexample — 80 years of nuclear non-use — involved a danger that was instantaneous, imaginable, and threatened decision-makers personally. Climate, overshoot, and institutional decay are the opposite. This is not a counsel of despair; it is a statement of the actual activation energy required, which is why my scenarios weight "crisis-driven mobilization in the 2030s" so heavily — that is historically how the species changes course, and it implies significant damage before the turn.

On interconnection specifically: Norman Angell argued in 1910 that European economies were too interdependent for war to be rational. He was right about the irrationality and wrong about the implication — 1914 happened anyway, because interdependence without trust is just mutual hostage-taking, and hostages get shot. "Too interconnected to fail" is not a stabilizer; it is the mechanism by which one failure becomes everyone's failure. The Late Bronze Age collapse — the best ancient analog — was precisely a trade-network cascade: palatial economies so interdependent that when the first nodes fell, the rest had no capacity to continue alone.

On collapse geography: every past collapse was regional, and recovery came from the outside — intact neighbors, trade partners, conquerors. A globalized civilization has no outside. This is the single strongest argument that a full cascade would be qualitatively different from Rome or the Maya. (Partial counterweight: knowledge is now cheap and ubiquitously stored, unlike in 1200 BC; but re-industrialization without easy fossil fuels has never been demonstrated, salvage and renewables notwithstanding.)

On prior collapses at the periphery: note honestly that by your definition, collapse has already begun — Syria, Yemen, Sudan, Haiti, Afghanistan, Libya, Lebanon, Venezuela have undergone exactly the process you describe: loss of complexity, safety, food, services, over years. Peripheral collapse is historically normal and does not by itself topple the core. The civilizational question is whether the failures reach the connecting tissue — food trade, finance, chips, energy flows — and the core states. So far, the core has absorbed peripheral failures while degrading its own buffers to do so.

The political economy of non-action — the actual crux

You are right that no electorate votes for less, and this is the load-bearing wall of the whole problem. But the precise mechanism matters, because it determines whether the wall can be moved.

Citizens have accepted less — repeatedly and with high compliance: wartime rationing, 90%+ top marginal tax rates for a generation after 1945, postwar European austerity. The conditions were: a visible shared threat, perceived fairness of burden-sharing, trusted leadership, and a sense of common fate. All four conditions are currently absent — and their absence is manufactured daily by the epistemic collapse you identified. The constraint is therefore not human nature; it is that the trust substrate required for collective sacrifice has been strip-mined — partly by bad actors, partly earned by elite behavior (impunity after 2008, Iraq, revolving doors, COVID whiplash). Distrust with rational foundations cannot be fixed by messaging; it requires sustained, verifiable elite accountability, which current power structures select against. This is a genuine trap, and I will not pretend to see the exit.

Given that, the historically literate forecast for the political path is:

1. Denial and delay (current phase, mostly).
2. Crisis improvisation under emergency powers when shocks cluster — sometimes impressive (wartime mobilization, an 11-month vaccine), but consuming fiscal and social capital each time.
3. Authoritarian drift, because allocation of scarcity by power is the default human solution when allocation by price or consent fails. The far-right rise you cite is the early phase of this, and note the loop: climate migration → destination politics hardens → origin countries destabilize further → more migration. The countries that demographically need migrants are politically closing to them; the incoherence is structural, not accidental.
4. Allocation conflicts between states — food export bans (2008, 2010, 2022 — each time contagious), export-land dynamics in oil (exporters keeping reserves, as you said — this has already happened to Indonesia, Egypt, the UK, Mexico, and it preceded every oil crisis), water disputes, and ultimately the 1930s pattern: economic stress + polarization + weak international institutions → war.

The only political path that has ever delivered large change without first requiring catastrophe is substitution, not austerity — and here the optimists have one genuinely strong card. Solar is now the cheapest energy in human history; batteries and EVs are on steep learning curves; China's industrial scale is forcing the transition at a pace no treaty achieved — Chinese emissions may be plateauing in the mid-2020s, driven by economics, not virtue. EVs are better cars; heat pumps are better heating. Where the transition offers "different but comparable or better," it proceeds without requiring sacrifice politics. Where it requires visible sacrifice, it stalls.

So the race is precisely this: exponential technology substitution versus exponential stress accumulation, with governance quality as the swing variable — and governance is currently degrading. That is the honest one-paragraph summary of the next twenty years.

The cascade anatomy — how the loops actually close

You asked for explicit interaction structure. These are the loops I consider load-bearing:

Loop 1 — The trust death spiral. Governance failure → earned distrust → non-compliance and populist capture → worse governance → deeper distrust. AI-generated content is an accelerant poured on this specific loop, because it destroys the evidentiary commons that any correction would require. This loop disables the response capacity for every other item on your list. I rank it the most dangerous meta-risk — more than climate physics, because climate is a Class B problem and this loop is what keeps it unsolved.

Loop 2 — Adaptive-capacity consumption. Each crisis is met with debt, emergency powers, and depleted stockpiles. None fully reconstitute. Fiscal space (global debt >330% of GDP; US interest costs now exceeding the defense budget), institutional trust, and political bandwidth are jointly drawn down. Meanwhile the claims on the remaining surplus — energy transition (~3–5% of global GDP/yr for decades), adaptation, infrastructure renewal, pensions and healthcare for aging populations, defense — sum to more than the surplus. That arithmetic is the "means diminish" mechanism made concrete, and it defines the window: the 2020s–early 2030s are the last period with the fiscal, energetic and demographic slack to execute transformation. After that, triage.

Loop 3 — Energy–finance–growth. Declining net energy (EROEI erosion as shale peaks late this decade, before substitutes fully scale) → structurally lower real growth → debt-based pension and finance systems become unsustainable (they are growth-dependent by design) → financial instability → underinvestment in the transition → worse energy position. A society can break this loop — but only by spending surplus during the transition, which is exactly what Loop 2 is consuming.

Loop 4 — Scarcity hoarding. Export bans in food, energy and minerals are individually rational and collectively catastrophic; they convert local shortage into global panic within weeks, and they are contagious, as 2022 demonstrated. The global food system feeds ~3+ billion people via trade; the Middle East imports over half its calories. The food system does not need to fail in absolute production terms to kill — it only needs the trade to seize. (Note the conditionality: trade is also our best buffer against local crop failure. It is simultaneously the resilience and the vulnerability, and which face it shows depends entirely on trust and geopolitics.)

Loop 5 — Maintenance debt. Under polycrisis, attention and budgets triage the urgent; the slow, invisible items (bridges, dams, nuclear decommissioning, munitions dumps, grid transformers) get deferred. They fail later, at the worst time, as shocks themselves. Your "aging infrastructure" and "legacy waste" items live here — individually second-tier, collectively a minefield being laid for a poorer future.

Stabilizing forces — stated with equal coldness. Crisis mobilization capacity is real (the species' one great historical trick). Technology cost curves are real and now partially decoupled from political will. Trade, if it holds, buffers local failures. Demographic decline eventually relieves throughput pressure. Knowledge is cheap, distributed and durable. Some adaptation is automatic where wealth exists. AI could compound scientific productivity in materials, catalysts, agriculture, fusion — genuinely. None of this is nothing. But note the asymmetry: most of the stabilizers are themselves dependent on the system holding (solar supply chains need the trade system; AI needs chips; mobilization needs trust), while the destabilizers are largely self-sustaining once triggered (ice loss, aquifer depletion, distrust, state failure). The virtuous loops are conditional; the vicious ones are not. That asymmetry is why the pessimistic scenarios carry more weight than a naive even-handedness would suggest.

Finally, the linchpin logic, which drives the headline probabilities. Preventing cascade requires simultaneous success on roughly five things: (1) no great-power war; (2) energy substitution outrunning depletion + climate damage; (3) the global food-trade system holding; (4) the financial system absorbing shocks without a sovereign doom loop; (5) a minimal epistemic commons sufficient for emergency collective action. Cascade requires failure on one or two. Even granting a generous 80% success probability per linchpin, the joint probability is 0.8⁵ ≈ 33%. This — not any single doomsday mechanism — is why prevention is the underdog even while every individual problem remains "solvable."

What "a chance" concretely means — and the signposts

Prevention is not a vibe; it is a list of conditions. The realistic version of "still a chance" looks like this: compound crises in the 2030s severe enough to break denial but not severe enough to break the system; mobilization on the 1940s pattern but sustained for decades; great-power détente driven by shared climate threat (there is Cold War precedent for cooperation under existential clarity); substitution technology reaching escape velocity before the fossil decline bite; and no nuclear use, no AMOC collapse, no unilateral geoengineering conflict in the interim. All achievable. None currently on trend. Perhaps one chance in three to four of assembling the sequence, declining each year of delay — because delay converts Class B problems into Class A and spends the surplus that the conversion requires.

Indicators that discriminate between scenarios, for anyone tracking this honestly: the AMOC fingerprint in North Atlantic temperature/salinity data; global grain stocks-to-use ratios and food export-restriction counts; US shale decline rates versus demand; the Taiwan Strait; insurance-withdrawal maps; AMR mortality trends; the count of failed states; trust indices; solar/battery deployment versus required trajectory; and any state moving toward unilateral stratospheric aerosol deployment — that last one will signal that the crisis phase has begun in earnest.

The answer, restated without cushioning

Is there still a chance to prevent the collapse of industrial civilization? Yes — a real one, roughly on the order of one in three, falling. The collapse is not written in physics; it is being chosen, daily, by default, through the historically normal mechanisms of elite insulation, electoral short-termism, security dilemmas and the strip-mining of trust. The species has the tools. The species has essentially never used tools of this scale before the rubble, and the buffer stocks that could absorb the rubble-phase are being spent in advance.

The most probable future is not the end of the world. It is a poorer, hotter, more violent, more fragmented world that keeps failing at the edges while the core degrades more slowly — a long emergency rather than an apocalypse — with perhaps a one-in-three chance that the failure reaches the connecting tissue and becomes genuinely civilizational, and a smaller chance that crisis finally triggers the mobilization that stability always required. The question "can collapse be prevented" is therefore already partly obsolete. The live questions are: how deep, how fast, how disorderly — and whether anything functional remains on the other side of the 2040s–2060s bottleneck. Those outcomes are still being decided. That, and nothing softer, is the truthful answer.