# Can a distributed thermal-electric fleet actually serve computation?

ID: ai-load-question · question · hypothesis · updated 2026-09-07
Tags: AI, grid, open problem, rationale

Data centres cluster at a few large sites with dedicated supply. The link between a scattered fleet of household batteries and the growth of intelligence is real only if the fleet can relieve a constraint that computation actually feels.

Three mechanisms are possible: freeing distribution capacity in the same service territory; providing a flexibility product a utility uses to approve new large loads faster; or directly following a computing site's timing signal. Each is testable. None is demonstrated. This is the most important open question in the rationale.

## Counterargument
The benefit may disappear when tested against a simpler alternative or under adverse operating conditions.

## Next test
Publish a bounded experiment, its baseline, its failure criteria and the observed outcome.

## Sources
- none

## Links
- challenges → is-this-good-for-ai
- depends-on → the-grid-is-full