“The next monetary regime turns on one variable — how fast AI capital becomes lower-cost output.”
The monetary thesis is now decisive: AI is tightening the economy before it is transforming it. The build-out is already pulling on power, compute, infrastructure and financing. The offset — higher output, lower unit costs and broader efficiency gains — is still arriving.
That means the first-order effect of the AI Supercycle is not disinflation. It is pressure on capital, inputs and rates. The sequence runs from AI CapEx to demand and input pressure, to a higher cost of capital, to less room for easing. Only later does AI diffusion drive higher output, lower unit costs and, finally, disinflation.
The evidence now lines up across three channels.
The Three Channels
Demand, inflation, supply response
-
AI CapEx flows into compute, power and infrastructure, lifting investment and financing demand, which translates into rate pressure. Williams notes that the aggregate efficiency effect is not yet clearly visible. signal: Capital demand runs ahead of any measurable efficiency offset.
-
Energy and input costs feed into prices, then into expectations and wages, forcing a monetary response. Bailey warns that policymakers may need to act before second-round effects are fully embedded. signal: Central banks may move ahead of embedded second-round effects.
-
AI adoption drives skills and job redesign, raising output per worker and lowering unit costs. Singapore is retraining more than 80,000 financial-sector workers to accelerate the transition. signal: The supply-side offset depends on how fast labor is redeployed.
Investor Implications
Positioning while the capital cycle leads
Europe's Test
Compress the chain
Four rules for the capital cycle
- Own the enablers selectively.
- Demand cash flow.
- Respect rate risk.
- Rotate when output gains become data.