Malaysia Data Center Bubble: Macroeconomic Cycles, Power Constraints, and the Coming AI Turning Point
Year: 2026
Type: preprint
Abstract: This paper examines the structural gap between contracted and utilized electricity capacity in Malaysia's data centre sector, drawing on Tenaga Nasional Berhad (TNB) and Suruhanjaya Tenaga (ST) data. As of December 2024, secured capacity across 38 projects totaled 5.9 gigawatts against actual operational load of approximately 405 megawatts — a utilization rate near 7%. The paper separates this verifiable infrastructure finding from a contested macro-cycle framing (Strauss–Howe generational theory, Dalio's Long-Term Debt Cycle) sometimes used to time AI-investment risk to 2026–2028, explicitly grading each claim's evidentiary basis. It documents an active tariff pass-through mechanism already used twice by TNB in under a year and offers regulatory recommendations distinguishing announced, committed, and financed capacity.
Author: Aodhan煜森, 定一居士
Language: English
Subfield: Information Systems
Field: Computer Science
Domain: Physical Sciences
Open Access status: green
OpenAlex ID: W7171565501
Cites:
Cited by:
Related to: