---
id: "concept-new-ai-triad"
type: "concept"
source_timestamps: ["§ Understanding the Constraints"]
source_url: "https://hbr.org/2025/10/is-ai-a-boom-or-a-bubble"
source_title: "Is AI a Boom or a Bubble?"
tags: ["infrastructure", "resource-scarcity", "scaling-laws", "ai-infrastructure"]
related: ["claim-physical-constraints", "action-secure-energy", "contrarian-physical-limits", "prereq-original-ai-triad", "action-workforce-partnerships", "question-energy-sustainability"]
definition: "The physical constraints of land, skilled trade labor, and energy that now limit the scaling of AI infrastructure, replacing the digital constraints of compute, data, and talent."
sources: ["futures"]
sourceVaultSlug: "hbr-seg-futures"
originDay: 2
articleStem: "hbr-foci-74-ai-boom-or-bubble"
sourceUrl: "https://hbr.org/2025/10/is-ai-a-boom-or-a-bubble"
sourceTitle: "Is AI a Boom or a Bubble?"
---
# The New AI Triad: Land, Labor, and Energy

Where the previous decade of AI was defined by a **digital triad — compute, data, and talent** (see [[prereq-original-ai-triad|The Original AI Triad]]) — the current scaling era is constrained by a **physical triad: land, labor, and energy**.

- **Land** — Data centers require massive plots with access to water and power, often facing local community resistance.
- **Labor** — The shortage has shifted from software engineers to **skilled physical trades**: electricians, construction workers, and technicians needed to build the infrastructure (see [[action-workforce-partnerships|trade-school partnerships]]).
- **Energy** — Identified as the **most severe bottleneck**, with projections that **U.S. data-center consumption will double by 2030** (see [[question-energy-sustainability|can sustainable energy scale?]] and [[action-secure-energy|secure long-term energy contracts]]).

The triad proves that while AI is a general-purpose technology, its underlying infrastructure is **not infinitely scalable** — the basis for [[claim-physical-constraints|the physical-constraints claim]] and the [[contrarian-physical-limits|contrarian insight that AI's limits are now physical, not digital]].

> **Enrichment note:** The energy leg is strongly corroborated — Goldman Sachs and other banks project U.S. data-center power demand roughly doubling by 2030. The land and skilled-trade legs are widely discussed qualitatively in infrastructure/real-estate literature; the *triad framing* is synthetic (the author's construction) but consistent with observed trends.


## Related across articles
- [[concept-ai-industrial-economics]]
- [[claim-ai-bottleneck-electricity]]
- [[concept-digital-sovereignty]]
- [[claim-energy-dictates-generative-ai]]