Research

A living map of the post-AI economy

When intelligence collapses in price, value doesn't stay in the intelligence layer. It leaves what can be reproduced, flows to the complements and the non-reproducible, and pools in categories that only become viable once cognition is nearly free. The essays follow it, with named companies, real figures, and the conditions that would make the argument wrong.

The time path Buildout · Digestion · Diffusion
Deployed capability Cost per unit of cognition BUILDOUT DIGESTION DIFFUSION YOU ARE HERE 2026 2025 2027 2030 2035
Phase I · now

Buildout

≈ 2025–27
What happens
The physical substrate goes up faster than it can be supplied. Power, grid, cooling and silicon become the binding constraints, and the constraint sets the price.
Where the value sits
With whoever owns the bottleneck. Scarcity earns the rent, not the software.
Rhymes with
  • Canal Mania, Britain1790–93
  • US railroad boom1866–73
  • Long-haul fibre buildout1996–2000
Phase II · next

Digestion

≈ 2027–30
What happens
The question turns from "can it be built" to "did it pay." Returns on the buildout are tested in public and financing gets selective. This is where the reset lives.
Where the value sits
With the survivors, and with whoever still holds cash. Most of the durable positions of the next decade are bought here.
Rhymes with
  • Canal share collapse1793–97
  • Railroad panic, then six lean years1873–79
  • Telecom bust2000–02
Phase III · later

Diffusion

≈ 2030–35
What happens
The capability stops being the story and becomes the substrate. It shows up in ordinary companies' margins rather than in vendors' revenue.
Where the value sits
With rails, trust, land and attention, and with the categories that exist only because cognition got cheap.
Rhymes with
  • Cheap bulk transport feeds the factory economy
  • National distribution, then mail-order retail
  • Dark fibre becomes broadband, cloud, streaming

Each phase pays a different owner. Getting the sequence wrong is the most expensive way to be right about the destination.

What the essays go after
  1. Which complements cheap intelligence consumes, and how long their scarcity actually holds before supply responds.
  2. What intelligence cannot reproduce, and what happens to the price of it when everything else gets cheap.
  3. Which businesses only become possible once cognition is nearly free, and which of them can hold a moat.
  4. Where control survives the collapse, and where the moat is already gone and the market hasn't noticed.
  5. What would make all of this wrong: a capability plateau, adoption that stalls, or the labs capturing the complements themselves.

The order changes as the evidence arrives. The questions are the fixed part.

Subscribe

Read it before it's consensus

Essays as they ship, a short note at every monthly close, and a full review each quarter.

Keep going
Investments

The same map, held as positions and marked to market every month.

See the strategy
Advisory

The same read, applied to the decision on your desk.

Explore advisory
Home

Back to the map: the four routes value takes.

Back to the map