Data Centers and Firm Power

Compute campuses want power that never stops; matching that need against an unbuilt machine requires unusual candor.

What a data center actually needs

A hyperscale data center is among the most demanding electricity customers that exists. It wants power that is continuous, high-quality, and extraordinarily reliable, and it increasingly wants that power to be low-carbon to meet emissions commitments. Firm clean power is a near-perfect description of the ideal supply — and a description that no fusion machine yet meets.

The burner housing MetroVolt is designed around this customer. It uses direct energy conversion of a D-3He plasma and carries a low neutron fraction of 5.44%, which is attractive for a generator sited close to sensitive equipment and people. Those are design properties of a simulation study, stated as such.

The reliability gap, stated plainly

The hardest requirement is uptime. Hyperscale Tier III operations target availability around 0.99982. The burner's studied availability sits in the range 0.86 to 0.995 — roughly 30 to 100 times more downtime than a Tier III facility tolerates. On current analysis, the burner is not a sole Tier-III-grade source. That is a physics-and-engineering finding, not a marketing footnote.

• Firm, clean, sited near load: the right shape of supply
• Low neutron fraction (5.44%): favorable for co-siting
• Availability 0.86–0.995 vs Tier III 0.99982: the open gap
• Best case near term: one firm source in a diversified supply, not the only one

The honest role for an early burner near compute is as a contributor to a diversified firm-power mix, not as the single source a data center bets its uptime on. See availability and hyperscale reliability and MetroVolt and data centers, honestly.