Distribute AI model training and inference across a global network of GPU providers. Verifiable, censorship-resistant, and radically cheaper than centralized cloud AI.
A complete ecosystem for decentralized AI — from training to verification to deployment.
Split large-scale AI model training across thousands of GPU nodes worldwide. XDNG's orchestrator handles data partitioning, gradient synchronization, and fault tolerance — so you can train models that are too large or too expensive for any single cloud provider. Pay only for compute used, with automatic checkpointing and elastic scaling.
Bid-based marketplace connecting AI model users with GPU providers for real-time inference at fraction of cloud costs.
Cryptographic proof system ensuring AI models were trained on claimed data with verifiable provenance and integrity.
GPU providers stake XDNG tokens to join the network. Higher stakes earn priority compute assignments and better rewards.
Novel consensus mechanism that verifies computation results without re-executing them, slashing verification costs by 99%.
XDNG token distribution designed for long-term network sustainability and decentralization.
350M XDNG allocated to GPU providers over 10 years
200M XDNG, 4-year vest with 1-year cliff
150M XDNG for grants, partnerships, and developer tools
300M XDNG across seed, strategic, and public rounds
Our path to a fully decentralized AI computing infrastructure.
Mainnet launch of XDNG compute network with initial 5,000 GPU providers. Token generation event and public distribution.
Launch of the distributed training orchestrator supporting models up to 70B parameters across multiple GPU clusters.
Real-time inference marketplace goes live with the neural proof verification system ensuring computation integrity.
AI model provenance tracking system deployed. Decentralized governance DAO activated for protocol upgrades.
Complete transition to decentralized governance. Network supports 100,000+ GPU providers and autonomous model marketplace.