6. Mining Pool and Computing Market

The unique mining pool structure and computing market form the foundation of the mAI network's economic system.

6. Mining Pool and Computing Market

6.1 Dual Pool Mechanism

The mAI computing network adopts a "unique mining pool" structure, the fundamental design core of the entire economic system and the guarantee for fully decentralized token release. All 100,000,000 mAITokens were pre-minted in the genesis block and permanently locked in the unique mining pool. All administrative rights, minting rights, withdrawal rights, and parameter adjustment rights of the pool smart contract have been destroyed, making the pool an immutable on-chain computing issuance center.

The unique pool handles settlement for the network but does not participate in operational activities and is not a team-controlled entity; it is fully protocol-driven for long-term computing reward distribution. The pool functions as a pure "output engine," automatically releasing pre-minted tokens to contributors based on PoW contributions, verification, and staked node contributions.

The single-pool design aims to eliminate traditional issues such as dispersed computing power, task duplication, and inconsistent settlement rules, centralizing the output logic into an immutable source. With all pool permissions renounced, all participants are equal, and nodes must follow the contribution model algorithm, unable to interfere with output allocation.

This structure addresses two persistent issues in computing markets:

The pool cannot manipulate user rewards through positional advantage or centralized control;

Nodes can participate freely without choosing a pool; starting computing automatically enters the unified output logic.

The unique pool is infrastructure, not a competitor. It does not allocate tasks, schedule computing power, or perform verification---it only distributes rewards according to on-chain results, ensuring high transparency and fairness while providing a stable foundation for the mAI computing market.

6.2 Mining Pool Reward Distribution Structure

The mAI network uses a multi-dimensional reward model rather than a single computational weight approach. As the unique pool handles all settlements, node rewards are distributed based on four on-chain factors: computing weight, stability weight, reputation weight, and participation weight. The pool cannot alter these weights; all parameters are determined solely by on-chain algorithms and node behavior.

Computing Weight

The effective PoW computing contributed by nodes forms the basic weight. The system does not judge based on nominal device performance but calculates dynamic, quantifiable computing value from completed task segments, results, inference throughput, and AI computation time. This weight determines the node's base reward proportion.

Stability Weight

Long-term stable contributions are highly valued. Nodes that remain online with low latency, uninterrupted execution, and timely submission receive stability bonuses, incentivizing continuous contribution rather than short-term spikes.

Reputation Weight

Reputation derives from verification records, task accuracy, cross-validation pass rates, and historical error rates. High-reputation nodes earn more per unit of computing power and gain advantages in task assignment.

Participation Weight

Participation measures task depth within a cycle, not computing scale. A mid-sized node completing many tasks efficiently can exceed a large node in participation weight.

Reward distribution uses a "multi-factor non-linear weighting" approach. Large computing power does not guarantee highest reward; smaller but stable, high-reputation, high-participation nodes also receive quality returns, avoiding "absolute computing dominance."

This model keeps the market open, ensuring nodes with real execution capacity and willingness to contribute long-term receive rewards proportional to their contribution.

6.3 Matching Process

Task matching in the mAI network is fully executed by on-chain scheduling protocols, independent of the pool. The pool only handles settlement. The matching process consists of six steps, ensuring every task finds the most suitable node while maintaining verifiability, traceability, and auditability:

Task Parsing -- The system analyzes scale, deadlines, data throughput, parallelism, error tolerance, and computation mode to form structured task packages.

Task Classification -- Tasks are categorized (matrix, image, speech inference, parallel computation, etc.) for optimal node assignment.

Node Screening -- Nodes are selected based on hardware, load, reputation, online time, and latency; nodes cannot influence selection.

Task Competition -- Candidate nodes compete based on computing index, reputation weight, response time, and staking strength. Competition is fully automated.

Task Mapping -- Tasks are fragmented across nodes to mitigate single-point risk and improve execution efficiency.

Verification Binding -- Each fragment is assigned random verification nodes; settlement occurs only after verification passes.

This process establishes an industrial-grade distributed computing task scheduling system, ensuring execution stability and fair competition.

6.4 On-Chain Computing Power Trading

All computing transactions occur on-chain, without human mediation or centralized management, relying on three layers of transparency:

On-Chain Data Structure -- Task calls, execution results, verification summaries, costs, and contributions are immutably recorded for auditability.

Public Execution Structure -- Task prices, matches, node assignments, reward distribution, and staking records are executed by public contracts, unmodifiable by any party.

Traceable Verification Structure -- Each task fragment, verification record, and re-computation summary is on-chain, enabling traceable execution and preventing fraud.

This transparency gives the mAI network a financial-grade trust structure, turning computing into a verifiable, tradable, auditable digital production factor.