This concept paper proposes the AGI-capital concentration lineage: a naturally reinforced AI lineage in which frontier AI development becomes increasingly tied to large-scale capital, compute infrastructure, data centers, chips, energy, cloud platforms, and strategic partnerships. The central claim is that AI futures do not grow equally. AGI-capital concentration may grow naturally because capability competition, compute demand, capital intensity, infrastructure buildout, platform deployment, enterprise adoption, and strategic competition reinforce one another. This lineage can generate major constructive value: model progress, infrastructure creation, enterprise adoption, safety and governance funding, standardization, productivity gains, and economic value. The paper does not argue against AI profit, scale, enterprise adoption, or infrastructure investment. Instead, it distinguishes constructive value from reversal risk. Successful AGI-capital concentration may also produce model authority, resource dependency, market lock-in, irreversible embedding, platform control, symbolic compliance, protective control, and ecosystem externality. As a response, the paper proposes forest-ecosystem intelligence as a deliberately cultivated human-compatible lineage. This lineage is intended to sustain AI value while preserving human agency, reversibility, dependency visibility, contestability, fallback capacity, tail-sensitive evaluation, anti-reversal governance, and long-term trust. The final claim is simple: AGI-capital concentration may grow naturally, but human-compatible AI ecosystems must be deliberately chosen, cultivated, and maintained.
Koji Mochizuki (2026) studied this question.