The Carbon–Silicon Symbiosis Equilibrium Regularity in Closed Planetary Civilizations: A Quantitative Validation Based on Century- Scale Simulation of a Million-Settler Settlement

The long-term co-evolutionary regularity of carbon-based and silicon-based intelligence in a closed planetary system holds urgency for both Mars settlement and Earth-side AI governance, yet existing research lacks cross-scale quantitative criteria. This study, built on the multi-scenario empirically validated Star Project ES-V2.0 kernel, constructs a century-scale (120-year) system-dynamics simulation of a one-million-settler settlement, covering a 32-run matrix formed by 6 governance scenarios × 4 supply-shock levels and 1,160 Monte Carlo envelope runs. The core finding is the Carbon–Silicon Symbiosis Equilibrium Regularity (SSE-R): the carbon–silicon collaboration ratio φ ≈ 0.60 constitutes a Pareto optimum, the only interval in the present parameter space simultaneously satisfying the dual objectives of "long-term internal stability" and "external expansion redundancy"; carbon-dominated regimes (φ < 0.50) exhibit, at the century scale, production stagnation and a civilization confined to its home soil, while silicon- dominated regimes (φ > 0.80) exhibit irreversible decay of human–machine consensus and elevated risk of chaotic collapse. A supply shock of−25% constitutes a systemic phase-transition inflection point, and under a strong shock (−50%) all governance paths collapse irreversibly. Through cross-scale comparison across four scales—the voyage segment (6 persons/9 months), thousand-settler settlement (1,000 persons/10 years), ten-thousand-settler town (10,000 persons/60 years), and million-settler settlement (1,000,000 persons/120 years)—the equilibrium point converges toward φ ≈ 0.60 as scale increases, preliminarily indicating the cross-scale convergence of this regularity. This study provides quantitative decision support for Mars settlement engineering and a time- scale-compressed reference framework for Earth-side AI governance.