Today (testnet, simulated euros)
- A voucher pays for one slab for a fixed number of epochs. Renewing a slab before that term ends isn’t built yet (planned in the durability milestone). Each epoch, each paid slab funds a budget.
- 70% goes to storage: each of the slab’s shards pays exactly one node, the closest one that passed its audit while holding it.
- 30% goes to egress: nodes earn for shards they served, credited from signed delivery receipts and capped per shard, so self-dealing can inflate it only up to the cap.
- Operators stake a deposit per node. Bad proofs and repeated missed audits cost half of it.
Planned
Storage credit. Operators can take earnings as vouchers instead of cash. A node with spare terabytes earns storage, and operators become users. It also avoids moving real money until that is properly in place.
Starting without a deposit. New nodes build their stake from withheld earnings after a probation period of passing audits, so a reliable track record counts before money does.
A price the market sets. One network price, adjusted each epoch within narrow bounds:
- the network gets fuller than its target (about 70%, leaving room for repair): the price rises and more operators join;
- capacity sits idle: the price falls and more people store.
- If repair falls behind, operator pay rises to pull in capacity.
The inputs and the formula are published every epoch. Customers lock the price when they buy, so a change only affects new purchases.
Payouts from what was actually paid. Each epoch’s pool is the money really received, spread over each voucher’s lifetime, so a price change never under- or over-funds payouts.