Headline finding. The single most important structural fact of the past month is a reversal I did not weight heavily enough thirty days ago: the Middle East de-escalation I judged to be 'materializing early' has unwound. Brent has jumped roughly 26% in a month, from $71.88 to above $90, as the US-Iran ceasefire deteriorated and President Trump threatened Iran directly; gold holds near $4,050 and the US 30-year yield sits at 5.21%. Equity volatility, meanwhile, is asleep at a VIX of 16. That bifurcation — anxious commodities and rates against complacent equities — is the emotional signature of the whole decade: a system that keeps absorbing shocks at the surface while its structural stresses compound underneath. The lesson I carry forward is humility about near-term junctures. Leaders with high behavioral volatility (Trump at BVI 8, cyclic escalate-then-declare-victory) make the oil and geopolitical tail fatter and choppier than any clean fork tree suggests.

State of the world — where we sit in the cycles. On the Kondratiev/technology cycle we remain in late installation, not deployment: capex and hype dominate, productivity does not yet. The RAG analogies are unambiguous — electrification ran 40 years (1882-1920s) from first plant to productivity boom, and the internet took a decade and a crash (1994-2005) before its gains showed in GDP. AI will differ on software-replication speed, but the base rate for general-purpose technologies delivering promised gains within a decade is roughly one-in-two, and I anchor to that. On the debt cycle we are at the peak, not yet in deleveraging: advanced-economy balance sheets are still expanding, China's contraction is sectoral rather than systemic, and the monetary regime is transitional. On the hegemonic cycle we are, on the British-American template, near the 1900-1905 equivalent — the challenger has surpassed the incumbent on one metric (manufacturing/PPP GDP) but has not converted it into systemic leadership, and full transitions take 40-60 years even after parity. On the demographic cycle the divergence is now the master variable: aging North and East Asia versus youth-bulge South Asia and Africa, with the US retaining the migration advantage and Russia in a historically unprecedented war-plus-emigration-plus-sub-1.5-TFR trap. What changed since last month is narrow but real: the oil/inflation tail re-fattened, and the Russian fiscal clock — which needs $90+ oil to run — was paradoxically given a reprieve by the very escalation that threatens everyone else.

Master scenarios. I construct four coherent worlds. Attritional Multipolarity (0.37) is the base case and the one requiring the fewest heroic assumptions: high real rates and aging populations cap the fiscal space for decisive assertion, so competition stays transactional; AI accrues as a metered utility without TFP escape velocity; swing states (India, Saudi, Turkey, Brazil) extract rents rather than picking sides; climate drifts on SSP2-4.5. Nothing breaks in any single quarter, which is precisely its danger. Fiscal Dominance & Sovereign-Stack Fragmentation (0.26) is where the debt-inflation-decoupling nexus resolves toward repression rather than deflation — the 1945-51 template made base case. I hold the economist's fiscal-dominance branch above consensus (0.20 vs 0.12-0.15) because democracies with aging electorates do not run primary surpluses and central-bank independence is politically besieged; the current 5.21% long yield against sub-2% growth is the tell. What makes it a *world* is the reinforcing geometry the analysts saw only in pieces: duplicated sovereign tech stacks raise the cost structure that fiscal dominance monetizes; closed borders tighten labor into the same inflation; and unaffordable conventional forces push powers toward cheaper nuclear hedging as arms control lapses. Contested Renewal (0.21) is the constructive world — AI productivity broadens above 2% in the US and India between 2028 and 2031 *after* a capex shakeout leaves cheap compute and cheap clean power behind (the dot-com fiber-glut rhyme). I keep it deliberately below a quarter because the productivity base rate is only ~50% and electrification's lag is a warning as much as a template. Systemic Confrontation (0.16) is the Taiwan-window rupture — most plausibly a blockade or quarantine during the 2027-2029 PLA advantage peak, stagflationary via a semiconductor/energy shock rather than deflationary, compounded by a Russian custody tail. I resist pushing it higher: nuclear deterrence is the largest structural departure from the classical Organski model, and the 75% transition-war frequency applies to the pre-nuclear class. Xi's profile (BVI 2, patient accumulation, economic coercion preferred, Taiwan tied to PLA readiness rather than a date) argues for coercion below the invasion threshold, which is why the military component is a parity-shock crisis cycle, not an invasion.

Cross-domain dynamics — the connections the analysts cannot see. Three linkages carry most of my synthetic weight. First, the compute-electricity nexus is the pivot between the base case and renewal: cheap clean power is the complementary input that converts AI capex into productivity (link_005) and simultaneously bends emissions — but in the utility-oligopoly path the same datacenter demand *absorbs* clean additions and slows fossil retirement (link_006). The identical technology can be climate's ally or its rival depending on whether the solar-battery S-curve outruns compute growth. Second, the fiscal-military coupling is underpriced everywhere: high debt does not just constrain butter, it reshapes guns — when conventional regeneration is unaffordable, states substitute toward nuclear deterrence precisely as New START lapses (link_007). Fiscal dominance is therefore a proliferation driver, not merely an economic regime. Third, the Russian fracture-to-custody channel (link_009): a war economy that cracks under a simultaneous external crisis is the single most systemically dangerous vector of the decade, and all three geopolitical council members converged on it being underweighted. Note the perverse feedback in the current data — the oil spike that threatens the West's inflation outlook is the same spike that funds Moscow's war economy past its breaking point, delaying the fracture while raising every other tail.

Critical junctures — what to watch. Five branch points structure the decade. The nearest is Iran/Hormuz (2026), live right now: renewed freeze (0.45), limited strike (0.38), or weaponization cascade (0.17). The Taiwan window (2027-2029) is the hinge between the base case and confrontation; deterrence holding at ~0.70 is my central assessment, with blockade/quarantine at ~0.22 the more dangerous-than-invasion middle path. The AI productivity inflection (2028-2031) decides renewal versus stagnation versus capex-reset; I put breakout at only ~0.24, modest gains at ~0.50. The Russian breaking point (2028) forks between adaptation (~0.55), face-saving freeze (~0.28), and fracture-plus-custody-risk (~0.17). And the sovereign-duration stress test (2027) — US 30Y above 5%, BOJ normalizing against >260% debt/GDP — decides whether the fiscal path is orderly or pivots to repression. Watch Brent's quarterly average, the 30Y yield, US productivity, PLA naval commissioning, and datacenter-clean-power matching as the five dials that will tell us which world is crystallizing.

Closing — calibrated confidence. I am most confident of the structural diagnosis: we are in late technology-installation, debt-cycle peak, and early hegemonic transition simultaneously, and these cycles are phase-locked in a way that produces compounding-but-managed stress as the modal outcome. I am moderately confident that no single master scenario deserves more than ~0.37, and that the confrontation tail, while real, is suppressed below its historical base rate by nuclear deterrence and US alliance density. I am least confident about timing and about the near-term geopolitical tail, where high-volatility leaders and a re-escalating oil market make any quarter capable of surprising me — as this month did. The Skeptic's central charge across every domain — that analyst scenario trees are cleaner than reality, with overlapping branches double-counting single pathways — is correct, and my response has been to fold those overlaps into integrated worlds rather than to pretend they are orthogonal futures. The honest summary: the next decade is most likely a hard grind, with a one-in-four chance it turns inflationary-and-fragmented, a one-in-five chance technology genuinely rescues it, and a one-in-six chance it ruptures. I assess, I do not predict.