When compute spend and return diverge
We are a research institution studying the economics of the AI build-out — and we study it from inside the transformation, not from the sidelines.
Capex Watch · live
Convergence flag active
Four of six fragility indicators elevated on the compute bellwether.
Open the board →
Founding teardown
The race between the lag and the runway
Is the build-out's spend justified before the financing reprices?
Read it →
The instrument
We measure ourselves
A real-time read on what AI absorbs — and what stays human.
See the read →
The thesis · two clocks
Every bubble is a question of whether the spending is justified by the output. We measure that question on two clocks at once.
Can the spending survive?
Capital expenditure versus demand. Circular financing. Depreciation that hasn't been marked. The arithmetic of whether the capex is paid for by revenue that actually exists.
Will the payoff arrive?
When does AI investment show up in aggregate productivity? Solow's computers took a decade to reach the statistics. We track the lag between adoption and measurable output.
Our entire edge is one measurement: the race between the productivity lag and the financing runway. The bull case is a lag-then-boom. The fragility case is capex repriced before the payoff ever lands — the lag outlasts the financing. No one else is timing both clocks. We can.
Flagship instrument
A live fragility scorecard for the AI infrastructure build-out: the divergence gauge, the circular-financing carousel, the depreciation exhibit, and the 2008 comparison — each claim sourced or labelled, every evidence gap named. The teardowns the consensus is skipping.
Our research is open right now — the live board, the exhibits, and the founding teardown. A gated subscription with ongoing papers is what it builds toward.
Explore the desk
Research you can read now — and a publication being built around it.
Research · gated
The live fragility scorecard and exhibits.
Research · open
The founding argument, with its falsifiers.
Research · the method
How we measure AI's economic absorption.
Publication · soon
Sourced news on AI economics.
Publication · live
Record results, record cuts — the cuts vs the capex.
Publication · live
The AI-stack tape — click any name for its fragility read.
The dual mandate
AI is both our engine and our subject.
We cannot effectively analyze the economics of AI from the sidelines. To understand its impact on productivity, labor, and capital allocation, we first master applying it inside our own operation. The result is research that is uniquely grounded — a think tank that is itself a live data point on the very shift it studies.
We don't just speculate on the future of the economy. We use the tools shaping it.
Beyond the published research, we run bespoke benchmarking — deploying our framework inside your operation to measure its true AI productivity yield against our baseline.