---
id: "claim-capex-obsolescence"
type: "claim"
source_timestamps: ["\\\"§ Inside the Firm", "the Same Fog\\\""]
tags: ["manufacturing", "robotics", "capex"]
related: ["concept-ai-fog", "action-stage-gate-capital"]
confidence: "medium"
testable: true
speakers: ["Toby E. Stuart"]
sources: ["futures"]
sourceVaultSlug: "hbr-seg-futures"
originDay: 2
articleStem: "hbr-foci-72-future-ai-fog"
sourceUrl: "https://hbr.org/2026/04/the-future-is-shrouded-in-an-ai-fog"
sourceTitle: "The Future Is Shrouded in an AI Fog"
---
# General-Purpose Robotics Will Collapse Specialized Manufacturing Setup Costs

**Claim:** Using a **hypothetical aerospace contract manufacturer**, Stuart argues that traditional **10-year ROIC models** for heavy CapEx — e.g., a **$30M CNC production line** — are fundamentally flawed in the AI era. A new generation of **AI-controlled, general-purpose robots** that *learn tolerances from sample runs* rather than months of setup will **collapse the setup costs** that historically protected specialized manufacturers, letting competitors or customers bring production in-house at a fraction of the cost. This is why stage-gated capital ([[action-stage-gate-capital]]) beats long-horizon commitment under the [[concept-ai-fog|AI fog]]; see also [[contrarian-corporate-planning]].

**Confidence: medium** (explicitly flagged as speculative). **Testable: yes** — via robotics setup-time and unit-cost benchmarks.

**Enrichment / verification:** A **forward-looking thought experiment**, not current practice. Adaptive CNC, collaborative robots, and ML-based quality control already cut setup time and enable flexible manufacturing, but they do **not yet fully collapse** setup costs in complex, high-precision aerospace production, where regulatory, safety, and quality constraints slow adoption. [[entity-waymo|Waymo]] is cited as physical-AI progress but is *transport, not manufacturing*. An expert would treat this as a **tail-risk scenario** for scenario planning, not an imminent inevitability.