Capacity For Hyperscale
Hyperscale campuses draw large, growing, continuous power; the burner is designed to supply firm base capacity at that scale through multi-unit plants.
The scale of the demand
A large AI campus is a substantial firm load, and new sites are being planned larger still. That demand is continuous and concentrated, which strains local grids and interconnection queues. Meeting it firmly requires a source that can be sited near the load and scaled by adding units rather than by stringing new transmission across a region.
The burner scales by multiplication: a plant is built from several units on a shared bus, adding firm capacity in blocks. This is also how the reliability target is met, since the same units provide the N+1 redundancy that a hyperscale load requires. Capacity and reliability are supplied by the same architectural choice.
Scale interacts sharply with two of the burner's gates. More units mean proportionally more helium-3, and a commercial unit already needs roughly 400 times current domestic He-3 supply, so a multi-unit hyperscale plant is firmly gated on a fuel supply that does not yet exist at terrestrial scale. This is why breeder-bred and lunar He-3 are part of the same story.
The honest scaling statement: the architecture can add firm capacity in blocks and reach hyperscale sizes on paper, but doing so at commercial scale depends on solving the fuel-supply and reliability gates first, with a test unit not expected before ~2032.
- Hyperscale load is large, growing, continuous
- Burner scales by adding units on a shared bus
- Same units provide capacity and N+1 redundancy
- Multi-unit scale multiplies the He-3 supply gate (~400x per unit)