Skip to content
adikumar.co
Research Frameworks Tools Advisory About
ESSAY ·The Investment Layer

The three downside cases

Three distinct scenarios in which the AI infrastructure investment map compresses, with transmission mechanisms, asset exposures, and signals to watch. Part VII of The Investment Layer.

16 August 2026 · 20 min read
ESSAY ·The Investment Layer

Underwriting AI infrastructure debt

The structures, spreads, covenants and rating criteria that determine how $200+ billion of AI infrastructure debt gets priced, and where the underwriting is soft. Part VI of The Investment Layer.

15 August 2026 · 20 min read
ESSAY ·The Investment Layer

The PE playbook, layer by layer

Five recognisable PE playbooks across the six AI infrastructure layers, their return profiles, and the three platform theses being constructed. Part V of The Investment Layer.

14 August 2026 · 21 min read
ESSAY ·The Investment Layer

The coverage asymmetry

Where equity research concentrates, where it misses the material stories, and what the coverage gap tells you about where mispricing sits. Part IV of The Investment Layer.

13 August 2026 · 20 min read
ESSAY ·The Investment Layer

The build-to-lease flip

Hyperscalers moved from majority-owned to majority-leased over seven years. The developer stack that made it possible, and where the return arithmetic breaks. Part III of The Investment Layer.

12 August 2026 · 20 min read
ESSAY ·The Investment Layer

The AI infrastructure M&A map

Every material deal 2023-2026 across the six layers of AI infrastructure, four through-lines that describe them, and fifteen forward candidates by name. Part II of The Investment Layer.

11 August 2026 · 22 min read
ESSAY ·The Investment Layer

How the AI buildout is financed

Six pools of capital finance the AI infrastructure buildout: hyperscaler cash flow, REITs, OEM balance sheets, structured debt, sovereign, and PE. Part I of The Investment Layer.

10 August 2026 · 23 min read
ESSAY ·AI Power Chain

The Services Inversion

An installed base generates maintenance work whether or not the manufacturer does it. Two conditions decide who captures the annuity: the customer's capability gap, and defensibility against independents. Seven positions across the AI infrastructure chain.

6 August 2026 · 34 min read
ESSAY ·AI Power Chain

Pricing Under Scarcity

Shortage is not the same as pricing power. Ten commercial positions across the AI infrastructure chain, sorted by the two conditions that determine whether scarcity converts to margin: how expensive escape is for the buyer, and whether qualified rivals decline to undercut.

5 August 2026 · 43 min read
ESSAY ·AI Power Chain

The AI Power Chain: Vendor Screen

300+ vendors across 6 stacks. Capacitors, wide-bandgap, thermal, interconnect, on-package delivery, modular DC. With a weighted diligence rank you can re-weight live. Filter by geography, size, public/private. Download as CSV.

4 August 2026 · 82 min read
ESSAY

The Modular Datacenter Stack Technical Companion

Skid architectures, BESS integration, MV switchgear, busway and rack topologies for the modular hyperscale datacenter build.

2 August 2026 · 16 min read
ESSAY

The Modular Datacenter Stack: How the Building Gets Built (When There Aren't Enough Electricians)

How the AI datacenter building gets built: modular skids, BESS, UPS, busway, rack enclosures. Where PE has deployed and where returns compress.

2 August 2026 · 31 min read
← Newer Page 7 of 9 Older →
Frameworks
  • The AI Power Chain
  • The Investment Layer
  • Due Diligence Framework
  • Grid-to-Chip
Series
  • The DC-DC Transition
  • The Regulatory Layer
  • Financing the AI Buildout
  • Tech Spotlights
  • Case Studies
  • Deal Tear-Down + Watch
  • What Changed (monthly recap)
Reference & Tools
  • Knowledge Base
  • Adi's Theses
  • Glossary
  • Vendor Screen
  • Question Banks
  • DD Evidence Matrix
  • Investment Scorecard
  • Investment Workbench
Engage
  • Advisory
  • About
  • adi@adikumar.co
  • LinkedIn
  • X (@NormalisedAlpha)
© 2026 Adi Kumar. All rights reserved. All content on this site — including essays, analysis, frameworks, charts, models, code, visualisations, and interactive tools — is the copyrighted work of Adi Kumar. No part may be reproduced, distributed, transmitted, republished, syndicated, adapted, translated, stored in a retrieval system, quoted at length, incorporated into commercial products or reports, or used to train machine-learning or artificial-intelligence models — in whole or in part, in any form or by any means — without the prior express written permission of the author. Brief quotations of up to 100 words with clear attribution and a direct link to the source page are permitted under fair use. All other uses require written consent: adi@adikumar.co.
Written by Adi Kumar · Privacy