Helium-3 Offtake and the Fuel Loop
Breeder helium-3 (~1.97 kg/yr) is both a saleable isotope and the burner's fuel; L7 arbitrates between external offtake and internal fleet demand.
A product that is also a fuel
Helium-3 is unusual in the Kronos material graph: the breeder produces it (~1.97 kg/yr class), and the burner consumes it as fuel in its D-3He reaction. So every kilogram is contested between external customers (quantum computing cryogenics, medical and national-security uses, neutron detection) and internal burner-fleet demand. L7 arbitrates this explicitly rather than letting one side silently starve the other.
Allocation as a constraint, not an afterthought
Helium-3 is stable (it does not decay), so unlike tritium the accounting is simpler in time but harder in allocation: it is genuinely scarce. L7 treats helium-3 as a shared pool with a policy that reserves the burner fleet's committed fuel need first, then releases the remainder to external offtake. The fleet dispatch optimizer reads the same pool as a hard fuel constraint, so a firm-power commitment is never made against helium-3 that has been sold externally.
# helium-3 allocation policy
pool = he3_inventory() + he3_breeder_inflow(horizon)
fleet_need = sum(burner_fuel_need(u, horizon) for u in burner_fleet)
reserved = min(pool, fleet_need) # burners first (firm power)
external = max(0.0, pool - reserved) # remainder to offtake
assert reserved + external <= pool # never over-commit
The scarcity of helium-3 is the reason the burner fleet's growth is fuel-gated. Breeder output alone supports a limited burner population; larger fleets depend on additional sourcing beyond the breeder foundry. L7 encodes this honestly: burner fleet plans are constrained by projected helium-3 supply and are never scaled beyond it in the dispatch and fleet-strategy layers.
External helium-3 shipments carry the same assay, custody, and byproduct-material record as tritium, and are drawn only from the released remainder. A customer sees firm availability only for quantities outside the reserved fleet pool.
Production and allocation run today on twin-projected breeder output at design stage. No helium-3 is delivered, internally or externally, before the breeder produces it at FOAK ~2030; the allocation machinery is validated now against the modeled pool.