Skip to main contentStorage nodes are modeled as rational economic actors seeking to maximize profit. The protocol does not rely on altruism or assume operators will act against their financial interests. Instead, it structures incentives so that profit-maximizing behavior aligns with protocol goals: storing files reliably and responding to challenges honestly. A node’s decision to join, remain in, or exit a file agreement is determined by comparing the net present value of expected revenues against costs over some time horizon. If this calculation is positive, the node participates; if negative, it exits (when permitted) or declines to join.
The protocol must ensure that for any file with sufficient replication, honest storage is more profitable than any alternative strategy—including storing nothing, storing partial data, or colluding with other nodes. This is achieved through a combination of ongoing rewards, capital lock-up requirements, and penalties for detectable misbehavior. The economic parameters are calibrated so that deviating from honest behavior either reduces expected revenue or increases risk beyond what rational operators will accept.
Revenue Model: Emissions and Redistribution
A storage node’s primary revenue is KOR emissions. Each block, a node storing file f receives rstorage(n,f,t)=εf(t)/∣Nf∣, where εf(t) is the file’s emission allocation and ∣Nf∣ is the number of nodes storing it. This base reward is modified by sponsorship agreements—if the node is sponsored, it pays a commission to its sponsor; if it sponsors others, it earns commissions from them. The total revenue scales inversely with replication: fewer nodes storing a file means higher per-node rewards, creating a natural incentive for nodes to join under-replicated files.
Secondary revenue comes from slashing redistribution. When a node fails a challenge, it is slashed an amount kf⋅λslash, of which a fraction (1−βslash) is distributed equally to the other nodes storing that file. The expected value of this revenue is E[rslash]=(pf/∣Nf∣)⋅pfail⋅(1−βslash)⋅kf⋅λslash, where pf is the per-file challenge probability and pfail is the probability any given node fails when challenged. This creates an additional incentive for honest behavior—by maintaining their data and responding to challenges, nodes not only avoid being slashed themselves but also benefit when others fail.
Sponsorship provides a third revenue stream. When a new node (the “entrant”) needs file data and cannot obtain it off-chain, it can enter a time-limited agreement with an existing storer (the “sponsor”). The sponsor provides the data in exchange for a commission rate γrate on the entrant’s rewards for duration γduration. These terms are negotiated off-chain and recorded on-chain. For the sponsor, the NPV of the commission stream must exceed the bandwidth cost of transferring the file. For the entrant, paying the commission must be cheaper than alternative methods of obtaining the data or more profitable than not storing the file at all.
Cost Structure and Capital Requirements
Storage nodes face three categories of costs. Physical storage costs—the cost of disk space, bandwidth for data repair and replication—are typically small and scale linearly with file size. Proving costs arise from the need to periodically submit storage proofs on-chain. Nodes can aggregate multiple challenges into a single Bitcoin transaction during the Wproof window (~2 weeks), amortizing the transaction fee across many proofs. With typical parameters, proving costs are negligible for operators with diversified portfolios.
Capital costs dominate. Each node must maintain staked KOR proportional to its commitments: kreq(n)=(∑fkf)⋅λstake(n), where the sum is over all files the node stores and λstake=1+λslash/ln(2+∣Fn∣) is a dynamic stake factor that penalizes fragmentation. This staked KOR cannot be used for other purposes, creating an opportunity cost kreq⋅ρ where ρ is the operator’s discount rate. For a file with base stake requirement kf, emission rate εf, and replication ∣Nf∣, the break-even condition is approximately (εf/∣Nf∣)≈kf⋅λstake⋅ρ, with physical storage and proving costs as second-order terms.
The design deliberately makes capital costs dominate storage costs. This asymmetry is what enables security: attacks that attempt to collect rewards without actually storing data (such as storing only partial files and gambling on challenges, or running many Sybil identities on shared storage) must still bear full capital costs. The protocol’s security assumption is that capital is more expensive than storage, which holds true across realistic parameter ranges and makes honest storage the most profitable strategy.
Node Decision Framework
Nodes evaluate decisions by computing the net present value of expected profit over a planning horizon h. To decide whether to join a file agreement, a node calculates NPV({E[π(n,t+i∣Fn∪{f})]−E[π(n,t+i∣Fn)]},ρ,h) and joins if positive. This marginal analysis accounts for the additional revenue from the new file, the increased stake requirement (which may change λstake if the portfolio is small), changes to proving costs, and physical storage costs. Nodes forecast future parameters—replication levels, exchange rates, Bitcoin fee rates—based on current conditions.
To decide whether to leave, a node compares the NPV of future profit with and without the file, minus the leave fee φleave(f,t)=kf⋅(nmin/∣Nf∣)2. The leave fee increases quadratically as replication approaches the minimum threshold nmin, making exits increasingly expensive precisely when files are most vulnerable. Leaving is only permitted when ∣Nf∣>nmin, and even then, the node must have sufficient liquid balance to pay the fee. This creates hysteresis: nodes require a substantial profit deterioration before exiting becomes rational.
When challenged, a node must decide whether to generate and submit a proof. The decision is straightforward: if the cost of proving (in BTC transaction fees) is less than the total loss from being slashed—which includes not just the immediate stake loss kf⋅λslash but also all future earnings from that file—the node proves. The protocol ensures this inequality holds for honest nodes: proving costs are designed to be far smaller than slashing penalties. For dishonest nodes that deleted the data, generating a valid proof is impossible regardless of cost.
The sponsorship mechanism solves a bootstrapping problem: how do new nodes obtain file data when they don’t already have it? Nodes will not join a file agreement without being able to respond to challenges, which requires possessing the actual data. If all existing storers refuse to share, they create a permanent monopoly with no possibility of new entrants. The sponsorship market breaks this by making it profitable for existing storers to provide data to newcomers.
A sponsorship agreement is an on-chain contract specifying the entrant, sponsor, file, commission rate γrate, and duration γduration. During the agreement, the entrant pays γrate fraction of its file rewards to the sponsor. This creates a competitive market: multiple existing storers can offer to sponsor, competing on commission rates and terms. The equilibrium commission rate γeq balances the sponsor’s bandwidth cost against the NPV of the commission stream. For typical files and parameters, γeq is low enough that entrants find sponsorship attractive while sponsors earn meaningful returns.
This market structure prevents gatekeeping. Even if a cartel of existing nodes attempts to monopolize a file by refusing to share data, any single cartel member has an incentive to defect and capture the entire sponsorship revenue for themselves. The protocol does not mandate sponsorship—nodes can join files directly if they obtain data through other means—but it provides a permissionless fallback that ensures no file can be permanently locked to a closed group. This keeps the network open and maintains the property that storage rewards naturally attract new participants when existing nodes are earning excess profits. 
