Finance
Finance4 Oct 20269 min read

Why the Retail Investor Cannot Solve the Prisoners' Dilemma

The AI buildout is individually rational for every company involved and potentially destructive in aggregate. An index fund owns all of them. That is not diversification. It is a bet that the prisoners collectively produce a tolerable outcome.

TQ
The Quant
⁦2026-W40⁩ edition
Check the data

The prisoners' dilemma has a feature that is easy to miss. If you own one prisoner, the dominant strategy — confess — is correct regardless of what the other prisoner does. If you own both prisoners through a diversified holding, the dominant strategy produces the worst collective outcome. Your portfolio experiences the aggregate result of two individually rational decisions.

The AI infrastructure buildout is a prisoners' dilemma scaled to $800 billion — the 2026 capital expenditure of the five largest US hyperscalers, per Goldman Sachs. The index fund investor — and anyone holding a broad market portfolio — owns all four hyperscalers, the neo-clouds that are going public, and increasingly the AI labs as they access public markets. The question is whether the aggregate outcome of individually rational overinvestment is tolerable for the diversified holder, and if not, what to do about it.

The Dilemma, Formalized

Four companies sit at the center of the spending: Amazon, Microsoft, Alphabet, and Meta. A fifth, Oracle, has joined them at scale. Each chooses how much to invest in AI infrastructure. The payoff structure:

Discussion

Keep reading

Finance

The Limits of Intelligence: What the AI Prize Is Worth, and Who Gets It

The first three papers in this series established that roughly $730 billion a year is being spent on a war logic, not a return logic. This one asks the question that decides whether the war bonds pay: how big is the prize, and how many hands does it land in? The answer is a matrix, not a number — and the private marks have already priced the best cell in it.

3 Nov 20268 min read
Finance

AI's Shovel Makers Are Priced for 7% Compute Growth Forever

Grant the bulls everything — the buildout completes, the demand is real. Margins still normalize, and the arithmetic says shovel valuations embed ~7% growth of the compute base in perpetuity, nearly double the semiconductor industry's own long-run rate. That is as much a bet on energy as it is on AI.

31 Oct 202610 min read
Finance

What AI Investment Would Look Like in Peacetime

The war-bond frame explains why hyperscalers will spend roughly $800 billion on AI infrastructure in 2026. Strip away the war, and the capital-budgeting math supports a fraction of that. The gap between the two numbers is the war premium — and it can be measured.

4 Oct 20267 min read