4. Computing Power Incentive System
The computing power incentive system ensures fair and transparent reward distribution based on computing contributions.
4. Computing Power Incentive System
4.1 Computing Contribution and Reward Calculation
Computing contribution is the sole and core basis for mAI rewards. All nodes' POW, AI task execution, and verification completion are recorded via on-chain verification, with the unique mining pool automatically releasing mAI according to unified rules. With the pool permanently renouncing authority, no individual or organization can alter algorithms or manually adjust distributions; rewards depend entirely on contribution and on-chain rules.
Contribution comprises three components: AI-PoW workload, task pass rate, and execution stability. AI-PoW, based on real AI tasks, carries the highest weight; task pass rate measures result validity; stability gauges continuous online performance.
Contributions are aggregated per cycle, producing on-chain contribution scores, and rewards are allocated proportional to network share. Records are immutable, ensuring auditable, computable returns. This model tightly links contributions to computing power, preventing waste and converting every computation into on-chain value.
4.2 Stake-Boosted Power Model
The Stake-Boosted Power mechanism applies to verification and high-value task nodes. Nodes stake mAI to gain eligibility for verification. Staking does not alter token issuance pace but affects task participation and reward proportion.
Staking has two incentive effects: first, it contributes to reputation, prioritizing nodes for high-value tasks; second, it serves as a participation threshold for validation. Only staked nodes can execute verification tasks and earn rewards.
Staking rewards come solely from the unique mining pool, executed automatically, with no adjustable parameters, ensuring absolute decentralization.
4.3 Mining Machine Levels and Yield Differentiation
Node levels are determined by real computing performance, execution efficiency, and stability, not hardware specs. Nodes earn dynamic on-chain computing weights based on long-term performance.
Scoring factors include hardware performance, historical success rate, task latency, and reputation. High-level nodes access more high-value tasks and higher rewards; low-performance or low-reputation nodes have reduced weights. This prevents hardware-dominance competition, encouraging stable, real workload completion, fostering a fair, healthy, and sustainable ecosystem. Rewards remain automatically distributed from the unique mining pool.
4.4 Mining Reward Pool and Node Incentive Pool Design
mAI has one mining pool with two algorithmic incentive layers: the mining reward pool and the node incentive pool. Both share the same 100M pre-minted tokens and zero-permission rules, differing only in release logic.
Mining reward pool: for AI-PoW tasks, rewards are allocated by contribution; no staking or node size distinctions.
Node incentive pool: rewards staked and verification nodes according to stake weight, reputation, and contribution; automatically released, with no human intervention.
Both pools draw from the same pre-minted tokens; rules automatically trigger releases. No adjustments are possible, ensuring strict decentralization.
4.5 Reward Decay and Output Balance
To maintain scarcity and sustainability, an on-chain reward decay mechanism is implemented. Algorithms adjust rewards automatically per period, with no permissions required.
Decay controls total issuance pace, maintaining balance between contribution and token circulation. For PoW, unit rewards decline as total computing power rises. For staked validators, increasing participants reduces individual rewards, balancing distribution. The sliding period calculation smooths reward curves, preventing volatility. All adjustments are on-chain, ensuring mAI scarcity is algorithmically guaranteed.
4.6 Mining Incentive Closed Loop
mAI forms a complete incentive loop: computing contributions drive AI-PoW output, staking drives verification ecology, together sustaining network growth. All rewards derive from the unique pool, which permanently renounces authority, making the system fully automated.
Loop logic:
Nodes contribute computing power; workload recorded on-chain
AI-PoW output automatically rewarded by the pool
Some nodes stake to participate in verification
Verification nodes execute tasks, earning additional rewards
All rewards distributed by the pool; no new tokens issued
Nodes continuously upgrade level based on contribution and reputation
Network computing capacity expands with participation
Expanded capacity enhances network task handling, attracting more demand