Working paper · CrossVol Research · Published 2026-06 · MPRA Paper No. 129363
Listed on MPRA, RePEc and OpenAlex under the title of the book it is adapted from: The China AI Disruption Thesis: Why the Sell-Side Is Six Months Late (deposited on MPRA on ).
This paper presents a structural derivatives framework for the repricing of the US AI infrastructure buildout. We formalize a nine-element analytical framework (five operational vectors and four geopolitical fronts) that converges on a 25-40% re-rating of pure-play AI infrastructure equities by Q1 2027. The five vectors are: (1) token commoditization, documented by the DeepSeek V4 Pro permanent pricing at 1/29th of frontier US output pricing; (2) Chinese hardware cost parity, with Huawei Ascend 910C achieving a 2.0-2.3x cost-per-performance advantage over NVIDIA H200 on inference workloads; (3) the US power grid bottleneck, with PJM 2026/2027 capacity auction clearing at $329.17/MW-day; (4) China's parallel energy buildout, with 12-24 month project timelines versus 4-7 years in the US; and (5) the hyperscaler bond wall, with $230-240B forecast 2026 issuance. We construct three trade families (equity dispersion, credit expression, and cross-asset hedges) and present an eleven-catalyst falsification calendar with probability-weighted thresholds.
This working paper is adapted from the book The China AI Disruption Thesis: Why The Sell-Side Is Six Months Late On AI Infrastructure (CrossVol Research, 2026).
BibTeX citation
@techreport{djouad2026airepricing,
title = {The AI Infrastructure Repricing: Five Vectors, Four Geopolitical Fronts, and the Structural Short},
author = {Djouad, Djellal},
year = {2026},
month = {6},
institution = {CrossVol Research},
type = {MPRA Paper},
number = {129363},
doi = {10.5281/zenodo.20509815},
url = {https://mpra.ub.uni-muenchen.de/129363/},
}