AI in 2026: Boom, Bubble, or Both? An Empirical Study of USD 750B in Big Tech CapEx (USD 545B Pure AI), Peer-Reviewed Workplace ROI, Accounting Depreciation Stretches, Jevons Paradox, and Startup Unit-Economic Failures
Year: 2026
Type: article
Abstract: IEEE Member Technical White Paper Series | 2026-11 | Working paper | Subtitle: Version 2.0 (Empirical & Academic White Paper) | Contributors: Sonu Kumar Singh (0009-0002-9180-4946) - https://orcid.org/0009-0002-9180-4946 | ResearchGate Preprint DOI: 10.13140/RG.2.2.17713.54882. Between 2023 and late 2026, the four largest U.S. tech companies (Amazon, Alphabet, Microsoft, and Meta) increased their combined annual capital spending from USD 141.2 billion to roughly USD 750 billion, with USD 520 billion to USD 550 billion per year going directly into pure AI infrastructure. This white paper evaluates four layers of the AI industry using SEC Form 10-K/10-Q filings (including 5-to-6-year server depreciation stretches vs. 3-to-4-year GPU economic life), Sequoia Capital's capital-recovery model, 1999-2000 telecom multiples, Jevons Paradox in token demand, cloud revenue round-tripping, and peer-reviewed workplace field experiments.
Author: Singh, Sonu Kumar
Institution: Capgemini (United States)
Language: English
Subfield: Artificial Intelligence
Field: Computer Science
Domain: Physical Sciences
Open Access status: green
OpenAlex ID: W7236178050
