Google and Amazon leverage custom chips for superior AI cost-efficiency. Google pursues aggressive infrastructure expansion to meet demand, while Amazon uses retail cash flow to sustain high capital intensity, solidifying data center dominance.

Key Highlights

• Custom Silicon Advantage: Both tech giants utilize custom chips to achieve better cost-efficiency and depreciation capabilities than competitors.
• Google's Expansion: Adopts a full-stack strategy, focusing on mega-campus expansions to address massive order backlogs.
• Amazon's Subsidy Strategy: Subsidizes high AI capital intensity using retail cash flow to establish superior data center density.
• Critical Success Factors: Future success depends on managing energy throughput and capital cycles effectively.

Table of Contents

1. Introduction
   • Google is Relatively Scattered in Deployment while Amazon Focuses on Northeast Region
2. Google’s Defensibility under Surging AI Capital Intensity upon Accumulating US$155 Billion Worth of Orders 
   • Google Gets Most of Its Revenue from AI, with AI Capital Intensity Still Rising
   • Google Broke Away from Convention by Announcing US$40 Billion of Investment in Texas
3. Gambling of the Retail Overlord: Dissecting the Dual-Sided Game behind Amazon's High AI Capital Intensity
   • Only a Small Portion of Amazon’s Revenue Is AI-Related, Yet Its AI Capex Intensity Remains Extremely High
   • Amazon's Long-Term Footprint in Virginia Gives It the Highest Data Center Density in North America Among the Four Major CSPs

<Total Pages:  9>