Where AI Value Lands

Proposed Structure

2 min read

Proposed Structure

~20 pages, English. Theoretical half (scenarios) + practical half (who wins, how), per the kickoff brief.

Note

Section titles below are working drafts. Names will mutate as hypotheses mutate.

Sections

  1. Executive summary (1pg) — the central question in one chart, the headline findings, the punch line.
  2. The stack — what an agentic AI system actually is (3pg) — working reference frame for the rest of the paper. Currently scaffolded by H1_L0-L7-ladder.
  3. Where the margin lands (4pg) — the economic question. Currently scaffolded by H2_u-curve-of-value. Historical analogies (Intel vs. PC OEMs, Stripe vs. acquirers, AWS vs. SaaS, Aggregation Theory) tested, not assumed. Counter-evidence included.
  4. Four scenarios for 2030 (4pg) — the theoretical half. Plausible futures with falsifiable signals to watch:
    • Hyperscaler consolidation — 3–4 frontier labs + clouds capture the stack.
    • Open-weight commoditization — Meta / DeepSeek / Mistral collapse the model moat.
    • Edge-first fragmentation — Apple-style on-device wins consumer and mid-market.
    • Sovereign AI fragmentation — geopolitics splits the stack along blocs.
  5. Geographic split — US / China / Europe (3pg) — same question, different optima per bloc. Different energy realities, regulatory regimes, capital pools. Currently scaffolded by divergent-value-stack-optima and geopolitics-global-ai-divide:
    • US — top-heavy optimum: hardware + foundation-model rent extraction; AGI-sprint framing critiqued by diffusion-vs-innovation.
    • China — bottom-heavy optimum: open-weight asymmetry (open-weight-asymmetry) deliberately commoditizes the model layer to push value into B2B integration (lee-01ai-pivot) and Global-South infrastructure.
    • Europe — regulatory-centric optimum: compliance (brussels-effect-and-mirage) + sovereign infra (sovereign-ai). Honest treatment of why this currently functions as handicap, not moat.
    • Note: Section 5 may need to function as a cross-cut of Section 3, not a standalone — see open question in 05_open-questions.
  6. The practical half — who wins, and how (4pg) — per scenario: which positions pay off, what the operator moves look like in practice, what the traps look like. Currently scaffolded by H3_orthogonal-axes-under-priced.
  7. 5–10 year visionary horizon (1pg) — energy as the long bottleneck, embodied / robotics, agent-to-agent economy. Speculative, clearly signposted.
  8. Appendix — link to the LLM-queryable wiki on GitHub, source bibliography, glossary.

Design notes

  • Visually strong infographics for the stack, the value map, the scenario matrix, the geographic split. Strong enough that screenshots get reposted.
  • Every strong claim either carries a citation or is explicitly labeled as a bet.
  • Each scenario is falsifiable: list signals that would confirm or deny it within 24 months.

Graph View

Backlinks