China: the highest absolute \(P_{in}\), racing its own \(\Phi\)

China is the largest single energy-harvesting node in human history. It is also the first such node to confront a \(\Phi\)-side cliff (demographics) before its \(\eta\) matures to first-world levels. The persistence ratio is robust today and ambiguous in 2040; the bet of the next two decades is whether \(\eta\) growth can outrun \(\Phi\) contraction.

What it is made of (\(\Phi\))

The substrate is enormous in capital, ageing in people, fragile in soil and water.

What it harvests (\(P_{in} \cdot \eta\))

• Primary energy: ~5.5 TW (~170 EJ/yr). About 29 % of world primary energy for ~17 % of world population. Largest single national \(P_{in}\) in history.
• Per-capita primary power: ~3,900 W — comparable to EU level.
• GDP: ~$18 T nominal, ~$33 T PPP. Second largest by nominal, first by PPP.
• \(\eta\) — improving steadily; energy-per-GDP has halved since 2005; still ~1.5× world frontier.

The energy mix is the single most consequential industrial fact of the 2020s:

• ~55 % coal
• ~18 % oil
• ~9 % gas
• ~18 % low-carbon (hydro + nuclear + wind + solar)

China is simultaneously the world’s largest emitter, the world’s largest renewable installer (~50 % of all solar and wind globally), the largest builder of new coal plants, the largest builder of new nuclear, and the largest EV manufacturer. The system runs every technology at scale. The bet is that the low-carbon share gets to ~50 % before peak emissions force a different conversation.

Bureaucratic complexity (\(\omega\)) and friction (\(\Gamma\))

\(\omega\) is unusual: superficially high (the Party plus the State plus the military all duplicate functions) but operationally lower than Western peers in time-to-deliver. A high-speed-rail mile in China costs ~$25 M and takes <2 years; a US equivalent costs $100 M+ and takes a decade. This is the single largest currently-observed \(\eta\) gap between China and the West.

\(\Gamma\) items, in order of severity:

• Demographic cliff — the dominant long-horizon item. By 2050 ~30 % of the population is over 65; pension and healthcare obligations were sized for a younger system.
• Real-estate / local-government-debt overhang — ~$8–12 T of overbuilt property + local government financing vehicles; not a Lehman moment, a Japan-1990-style slow grind.
• Taiwan — open file since 1949. The cost of not resolving rises; the cost of attempting to resolve militarily is enormous. Probably the most-modelled \(\Gamma\) item in any planning room on Earth.
• South China Sea — frozen with friction; tactical wins, strategic alienation of every coastal neighbour.
• Xinjiang / Tibet / Hong Kong — internal closure achieved at the cost of permanent external \(\mathcal{D}_{KL}\) on Chinese self-narrative (“peaceful rise”).
• US tech decoupling — semiconductors, AI compute, EUV, advanced packaging. Cost-of-substitution is high but not infinite; the workaround timeline is ~5–10 years per node.
• Belt and Road bad-debt — ~$1 T outstanding; many borrowers cannot repay; raises \(\mathcal{D}_{KL}\) on the foreign-policy ROI model.

What it believes about itself (\(\mathcal{D}_{KL}\))

China’s \(\mathcal{D}_{KL}\) has a distinctive structure: operationally low, narratively high. The system makes shockingly accurate forecasts at the operational and industrial-policy level — its five-year plans actually hit their physical-output targets at much higher rates than Western planning documents — and shockingly poor forecasts at the ideological and reputational level.

Delusion terms, ranked:

1. Information feedback under personalist control. Since ~2018, internal reporting has visibly degraded; bureaucrats know what Xi Jinping wants to hear and report accordingly. The 2022 zero-COVID exit decision had textbook \(\mathcal{D}_{KL}\) symptoms — model-vs-reality lag, sudden reversal under social pressure. This is the dominant long-term delusion risk.
2. Demographic optimism. Official models repeatedly under-forecast the pace of decline. Each census revises the prior downward; the trajectory is well below the official “soft landing.”
3. Taiwan-as-already-Chinese. Empirically the median Taiwanese identifies as Taiwanese, not Chinese, and the gap is widening generationally. Mainland strategy assumes the gap is reversible by economic and military pressure; observable evidence is the opposite.
4. “The East is rising, the West is declining.” A correct relative observation that becomes delusional if read as absolute rather than ratio. The West isn’t disappearing; the multipolarity story is correct, the unipolar-replacement story isn’t.
5. Wolf-warrior diplomacy effects. Repeatedly observed to harm Chinese soft power; repeatedly continued because it serves domestic legitimacy. Classic \(\mathcal{D}_{KL}\) trade — domestic short-term gain, international medium-term cost.

The Chinese system’s \(\mathcal{D}_{KL}\) is lower than Russia’s on most observable axes but trending in a worse direction because of the personalisation under Xi.

What shelters it (\(\Psi\))

• Trade openness — even amid decoupling, ~$3.5 T of merchandise exports; the world cannot rapidly replace.
• Russia, North Korea, Iran, Pakistan, large parts of Africa and Southeast Asia as quasi-clients, low individual weight but cumulative \(\Psi\).
• BRICS+ as a developing alternative-bloc framing.
• Internal market — the largest single consumer market by volume; can sustain industrial scale even under partial decoupling.
• Manufacturing supply-chain centrality — the world cannot quickly find another node that does what Shenzhen + Suzhou + Yangtze Delta do at price. This is sticky \(\Psi\) — falling but slow.

The largest threats to \(\Psi\) are US-led tech containment and the demographic clock.

What it actually keeps doing well

• Mass infrastructure deployment at price-quality nobody can match.
• Electric vehicles, solar PV, batteries, drones, ships — the cleantech of the 2020s and 2030s is being built in China at scale, often by Chinese firms ahead on cost, scale, and time.
• Poverty elimination (extreme poverty effectively at zero from ~400 M people in 1990, on World Bank’s $2.15/day line).
• Patent volume, AI publications, manufacturing PMI, port throughput — every leading industrial indicator runs through the country.
• Stable, low-violence society with a fraction of the social pathologies (opioid, gun, homelessness) of comparable-wealth Western peers.

Collapse trajectory if \(\mathcal{R}\) slips below 1

Not in the cards this decade. The plausible long-arc risk is Japan-1990 in slow motion — a long stagnation as the demographic and debt overhangs grind down growth — with a non-trivial Taiwan-conflict tail risk.

A 2027–2032 Taiwan miscalculation is the single decision that could compress \(\mathcal{R}^{(\text{China})}\) to dangerously low levels: a blockade or invasion attempt would trigger sanctions of unprecedented scope, mid-Pacific naval conflict with the US, and probable global recession. The persistence-ratio cost would dwarf any conceivable benefit to the node. The Party knows this; the question is whether \(\mathcal{D}_{KL}\) on Taiwanese resolve and US response remains low enough for the calculation to stay rational.

The lowest-\(\mathcal{D}_{KL}\) outcome for China: continue industrial \(\eta\) accumulation, manage the demographic glide path, keep Taiwan in a no-war indefinite-frozen state, lean into climate-tech export dominance. This is the rational path. The personalist regime makes it less than certain.

See also

• usa.md — the rival and the customer
• russia.md — the junior partner
• eu.md — the partner that is mid-decoupling
• united_nations.md — China is rising fastest in UN agency leadership