Speed To Power

Speed-to-power is how fast a campus can be energized to serve compute; on-site firm generation can shorten the path, but the burner's own timeline is honest and distant.

How fast can compute be powered

Speed-to-power is the time from deciding to build to energizing servers. It is set by generation, transmission, interconnection, and permitting together. On-site firm generation can shorten the transmission and interconnection parts by reducing dependence on new lines and on a full-requirements grid connection.

But candor requires stating the burner's own schedule plainly. MetroVolt is a design and simulation study. The program builds the breeder (Hyperion) first, with construction from Q2 2027 and first tritium near 2030. A test burner is targeted near 2032 and a commercial fleet near 2036. There is no burner speeding anyone's data center to power today.

So speed-to-power here is a future property, contingent on the burner maturing and on its physics gates being retired. What can be said now is architectural: once such units exist, generating firm power on site is a lever on speed-to-power because it decouples a campus from transmission build-out and from the slowest parts of interconnection.

No economic or schedule promise is made. The honest claim is that on-site firm generation is a plausible future contributor to speed-to-power, Any operator planning around MetroVolt should treat it as a source to design for in future builds, while meeting present demand with the mature parts of the hybrid it will eventually join.

• Speed-to-power spans generation, transmission, interconnection, permitting
• On-site generation can shorten the grid-dependent parts
• Burner is a design study: test ~2032, commercial ~2036
• Any speed benefit is a future, contingent property