AI Infrastructure Economics

A structured overview of how profit capture, capital burden, pricing power, and cycle risk are distributed across the AI infrastructure value chain.

Framework in development

This page frames the broader industry question behind the company profiles. The goal is not to list AI-related companies, but to understand how financial outcomes differ across value-chain positions.

Framing

Core thesis

AI infrastructure is a capital-intensive buildout where profit capture, capital burden, pricing power, and cycle risk are unevenly distributed across the value chain.

Company profiles provide the evidence layer; this page organizes those company-level metrics into an industry-level framework.

Structure

Value chain map

Cloud Platforms can be treated as future scope if current financial metrics are not yet collected.

Equipment

Companies

ASML

Key question

Which suppliers define the physical limits of advanced manufacturing?

Foundry

Companies

TSMC

Key question

Who carries manufacturing capital intensity?

Memory / HBM

Companies

MU

Key question

Is HBM a structural margin reset or an extension of the memory cycle?

Compute Platforms

Companies

NVDA, AMD, AVGO

Key question

Who captures accelerator economics and platform-level margins?

Networking

Companies

ANET, AVGO

Key question

How much AI cluster value flows into networking fabrics?

Power & Cooling

Companies

VRT

Key question

How does rack density translate into facility-level demand?

Cloud Platforms

Companies

MSFT, AMZN, GOOGL, META, ORCL

Key question

Who funds the AI infrastructure buildout, and how does CapEx convert into cash flow?

Questions

Core questions

Who captures margin?

Compare gross margin, operating margin, and operating leverage across value-chain positions.

Who carries capital burden?

Compare CapEx intensity, free cash flow pressure, and asset intensity.

Who has pricing power?

Use margin durability, ecosystem dependence, and supply constraints as initial signals.

Who absorbs cycle risk?

Track inventory, margin volatility, memory cycles, and demand normalization risk where data is available.

Coverage

Data coverage

Status reflects what exists in this analytics project today—not a claim that every name below already has collected financial metrics.

Completed
  • NVIDIA — company profile available
  • AMD — company profile available
Pending data collection
  • TSMC
  • Broadcom
  • Micron
  • ASML
  • Arista
  • Vertiv
Future scope
  • Microsoft
  • Amazon
  • Google
  • Meta
  • Oracle

Evidence

Company evidence nodes

Completed company profiles provide the evidence layer for future cross-company comparison views.

NVIDIA

Role

Compute platform

Signal

Rapid revenue expansion, high operating margin, and low direct CapEx intensity in the collected company profile.

Status

collected

AMD

Role

Compute challenger

Signal

Expanded revenue base, elevated R&D intensity, and materially lower operating margin than NVIDIA in the collected company profile.

Status

collected

Roadmap

Next build sequence

  1. Create first NVDA vs AMD compute economics comparison view
  2. Add TSMC collected financial metrics
  3. Build TSMC company profile
  4. Introduce foundry capital intensity into the comparison framework
  5. Expand toward hardware supply-chain comparison

This framework page is descriptive only. It does not recommend transactions in any security and does not present a finished comparison dashboard.