Gate Four in Depth: Availability Versus Tier III

Modelled burner availability of 0.86 to 0.995 is 30 to 100 times short of hyperscale Tier III uptime at 0.99982, so Aegis is not a sole Tier-III source today.

The uptime gap

Hyperscale data centers built to Tier III expect availability around 0.99982 — minutes of downtime per year. Modelled Aegis availability lands at 0.86–0.995. Expressed as downtime, that is 30–100x more outage than a Tier III facility tolerates. As a design study, Aegis cannot claim to be a lone Tier-III power source.

Why fusion availability is hard early

First-of-a-kind machines have unproven component lifetimes, maintenance intervals, and failure modes. The 5.44% neutron fraction, though low, still drives component wear that must be characterised before availability can be trusted. Availability is earned with operating hours, which do not exist yet.

What could close it, and the honest framing

• Redundancy and N+1 subsystem design
• Fleet operating hours that reveal and remove failure modes
• Aegis paired with grid or storage rather than sold as a sole Tier-III feed

The comparison to Tier III is deliberately unflattering because pretending otherwise would be dishonest. A first-of-a-kind machine has no operating hours, and availability is a property you measure over years, not one you specify on a drawing. The credible near-term role for Aegis is as resilient sovereign power paired with grid or storage, not as a lone Tier-III feed.

See the availability roadmap and the gate summary.