The China AI Disruption Thesis: Why The Sell-Side Is Six Months Late On AI Infrastructure

CrossVol Research, Boek 3 van 3 · Published

About this book

Why the sell side is six months late on AI infrastructure.

The sell-side consensus on US AI infrastructure (hyperscaler capex compounding at 25 to 30% a year through 2028, US power demand doubling by 2027, a memory supercycle) is, the book argues on quantifiable grounds, six months late. It documents five convergent shocks: token commoditization led by DeepSeek pricing, Chinese hardware reaching cost parity, the US power grid bottleneck, China's parallel energy buildout and the hyperscaler bond wall, with a four-front geopolitical layer (Iran, Greenland, Venezuela and Cuba) beneath them. The framework is built to be falsifiable: every catalyst is dated, every risk carries a probability, and a catalyst calendar runs from the third quarter of 2026 to the second quarter of 2027.

Intended readers. For derivatives strategists, equity research analysts, and cross-asset macro PMs.

Reeks CrossVol Research

  1. Boek 1 van 3: Beyond Gamma Exposure: A Four-Lens Framework for Options Traders Who See What GEX Misses
  2. Boek 2 van 3: FX Traders vs Brokers: Vanilla and Exotic Options, Forwards, and Other OTC Structures: What Retail Traders Never See
  3. Boek 3 van 3: The China AI Disruption Thesis: Why The Sell-Side Is Six Months Late On AI Infrastructure

CrossVol Research · Bekijk de reeks op Amazon

Working paper

The AI Infrastructure Repricing: Five Vectors, Four Geopolitical Fronts, and the Structural Short

MPRA 129363 · DOI 10.5281/zenodo.20509815

Buy on Amazon

Vertalingen

Identifiers

Read