Decentralized Network Trains 16B Model Across 3 Continents Using RTX 4090s
bittingthembits · x · 2026-07-29
The decentralized AI network $TAO has successfully trained the Orion-16B model across 256 heterogeneous GPUs, including RTX 4090s and 5090s, spanning three continents.
The project aggregates idle compute from independent global providers, coordinating machines with varying speeds and reliability to function as a single training system. With the training process publicly visible and contributors compensated for useful work, it demonstrates the viability of permissionless, decentralized compute networks for large-scale model training.
Related event: Decentralized Network Trains 16B Model Across Continents(4 posts)→
More from Infra
- DIY local AI server uses retired NVIDIA cards for about $165 total — blelbach · 2026-07-29
- Cradle Codec: GPU-Native KV Cache Compression Explained — knowrohit07 · 2026-07-29
- Cheap local intelligence could shift AI workloads away from the cloud — PeterDiamandis · 2026-07-29
- Bull case says AMD profit could 10x as AI spend and inference demand scale — AccBalanced · 2026-07-29
- Cradle Codec compresses KV cache for Ethernet transport between GPU nodes — knowrohit07 · 2026-07-29
- Bittensor raises q from 0.61 to 0.75, easing its emission gate for mid-ranked subnets — markjeffrey · 2026-07-29