As a capability scales, its binding constraint tends to leave the capability itself and settle on the physical resource that feeds it. What looked like a contest of the visible layer becomes a contest of the substrate underneath. The Philippine data center race illustrates it: the competition is nominally about compute and digital services, but the decisive factor for hyperscale investment is electricity — "stable, available and competitively priced." AI demand, "requiring massive continuous electrical consumption for servers and cooling," pushes the constraint down onto the grid until "the data center competition intensifies into a power infrastructure race." Whoever controls the substrate sets the ceiling on the capability. The practical move: name the true constraint before optimizing the visible one. Effort spent on the layer that is not binding buys nothing. ## Cross-Domain Connections - **Compute → power**: the source's case; the algorithmic race is now decided by kilowatt-hours and cooling. - **Commerce → logistics**: an online storefront's ceiling becomes fulfillment and shipping capacity, not the website. - **Careers**: past a skill threshold, advancement binds on relationships and distribution, not raw competence. - [[Find Each Component's Red-Line Number]] — the operational, per-component form: capacity is set by the first resource to hit its red line - Sibling: [[Parallel Infrastructure Planning Beats Sequential Buildout]] — how to build once the true constraint is named. - [[Infrastructure Repurposing for Home Networking]] — a concrete case: gigabit over old phone wire works because the binding constraint is the endpoint intelligence, not the physical medium, so you upgrade the substrate's smarts rather than replace the wire - [[Claude Inference Cost Problem - Brandon Gell X Thread]] — a corporate-strategy instance: Anthropic's competition with OpenAI looks like a model-quality contest, but the prediction is that it resolves on the substrate underneath — who owns the data centers — exactly as this framework predicts - [[Joseph Somsel NGCP Planning Report LNG Reliance Critique]] — the supply-side failure mode: PH grid planners are adding LNG and wind, not the reliable baseload the substrate needs, so the constraint this note names on the demand side goes unmet on the supply side too ## Source - [[Philippines Data Center Push Faces Power Reality Check]] — Philstar.com / Brix Lelis, June 28 2026 — https://www.philstar.com/business/2026/06/29/2538394/philippines-data-center-push-faces-power-reality-check