The Helium-3 Supply Gate

A commercial burner needs roughly 400x current domestic helium-3 supply per unit; a multi-unit data-center plant is firmly gated on a fuel that does not exist at terrestrial scale.

The fuel that isn't there yet

The burner runs on deuterium and helium-3. Deuterium is abundant, but helium-3 is not: a single commercial unit's demand is on the order of 400 times current domestic supply. A multi-unit data-center plant multiplies that further. This is a hard physical gate, stated as such, and it is not an economic statement.

This gate is why the Kronos program is structured the way it is. The breeder (Hyperion) is designed to produce helium-3 (about 1.97 kg/yr class) alongside tritium and neutrons, so the burner's fuel can be bred rather than mined from a supply that does not exist. Lunar helium-3 is the other long-horizon source. Both are why the breeder is built first, from Q2 2027, before a commercial burner fleet near 2036.

For a data center the implication is direct: no matter how well the availability and electrical architecture are solved, a commercial-scale MetroVolt plant cannot be fueled today. The fuel supply is a prerequisite, and it is being built deliberately through the breeder and, later, lunar sources. Presenting the burner as ready to power hyperscale campuses now would ignore this gate.

The honest position: the burner's electrical and physical case is real, but its commercial deployment is gated on a helium-3 supply chain that must be created — breeder-bred first, lunar later — and until then MetroVolt remains a design and simulation study.

• One commercial unit needs ~400x current domestic He-3 supply
• Multi-unit plants multiply the shortfall
• Breeder is designed to breed He-3 (~1.97 kg/yr class); lunar is the later source
• Commercial deployment is gated on a fuel supply that must be created