A trusted shared-compute ecosystem for AI training and inference
Corex Token turns distributed GPU, CPU, and edge resources into a verifiable on-chain compute network where every contribution becomes measurable economic value.
A proof path for foundational compute work tied to real hardware output, uptime, and task quality.
A higher-value proof path for model training and validation tasks tied to accuracy, efficiency, and stability.
Why the market needs a decentralized compute layer
AI workloads are scaling fast, yet centralized compute platforms still suffer from idle capacity, opaque scheduling, high participation barriers, and misaligned incentives. Corex Token rebuilds this value chain with verifiable rules.
Heterogeneous node coordination
GPU, CPU, FPGA, edge devices, and household hardware join the same global compute pool.
Task slicing and scheduling
Training and inference jobs are split into executable fragments and routed transparently across the network.
Contribution-driven rewards
Node work, staking, and validation all map directly to CXT incentives under public rules.
Reputation-weighted quality
Reliable long-term nodes earn better routing priority, higher-quality tasks, and stronger governance weight.
Data and model collaboration
The network supports workflow collaboration, model lifecycle visibility, and reusable training assets.
Multichain compute mobility
A compute bridge connects demand and resources across multiple on-chain ecosystems.
From task submission to verifiable settlement
Corex converts compute activity into economic proof through task slicing, Compute-PoC, Model-PoT, and on-chain settlement.
Task submission
A training or inference request enters the Corex network.
Node matching
The scheduler routes work based on hardware profile, reputation, and staking status.
Proof generation
Nodes execute assigned work and return verifiable proof for network-level validation.
Reward settlement
Smart contracts record work and distribute CXT according to transparent contribution rules.
const task = submitTask({ type: 'training', dataset: 'encrypted-batch' });
const assignment = scheduler.match(task, { stake, reputation, hardware });
const proof = node.executeAndProve(assignment);
const reward = settlement.distribute(proof, 'CXT');A full-stack ecosystem around production, validation, and circulation
Corex is more than a compute pool. It is an AI infrastructure layer built from node systems, task markets, staking enhancements, reputation logic, and multichain connectivity.
Node contribution and task allocation
Onboards nodes, maps their capabilities, and routes compute fragments dynamically.
AI training task marketplace
Matches training, inference, and validation workloads within the same settlement layer.
Node reputation system
Builds long-term credit from stability, completion quality, and contribution depth.
Stake enhancement model
Uses CXT staking to unlock premium task access and governance participation.
Data collaboration and model management
Supports coordinated workflows, dataset processing, and model lifecycle organization.
Multichain compute bridge
Connects external task demand and on-chain ecosystems to Corex supply.
CXT Token Model fixed, transparent, contribution-led
CXT is the economic layer for rewards, settlement, staking enhancement, and long-term governance inside the Corex network.
The total supply is permanently capped at 100,000,000 CXT.
Emission follows one public release path with no team reserve or private allocation.
Compute-PoC and Stake-to-Validate shape issuance and reward distribution together.
Token holders and staked nodes gradually gain proposal and parameter governance rights.
Governance and security discipline
Corex begins with foundation-led execution, operations, and security oversight, then evolves toward a CXT-centered DAO model.
Early network management
Technical roadmap, scheduling strategy, and parameter changes remain tightly coordinated during bootstrap.
Progressive decentralization
Governance authority expands over time as the network matures and participation deepens.
Security and compliance review
The system continuously monitors fake compute, abnormal behavior, and contract risk.
AI-native risk control
Automated risk models help score node reliability, detect anomalies, and adjust trust weights.
Scenarios across training, inference, and collaboration
Model training
Supports elastic GPU demand with task slicing, verifiable execution, and transparent settlement.
Inference services
Trusted node capacity absorbs high-concurrency inference requests while balancing performance, visibility, and cost.
Data collaboration
Connects data processing, model iteration, and node reputation into a long-term coordination workflow.
Start from the whitepaper and follow the full docs
Project mission, network design, ecosystem modules, tokenomics, governance, and risk management all continue in the docs. The homepage only acts as a guided entry.
Open /docs/introduction to continue with the Corex Token whitepaper