01
Resource layer
CPU, GPU, memory, storage, bandwidth, and availability contributed by independently operated machines.
GRID white paper · overview
Millions of machines.
One compute fabric.
GRID’s thesis is simple: the largest useful computer will emerge from connected machines working together. The network measures real resources, verifies completed work, and turns independent hardware into shared capacity.
10 pagesArchitectureProof of ResourceResource economy
The central idea
Useful computation is the work.
Traditional mining proves expenditure on a protocol-specific puzzle. GRID instead measures contribution to authorized workloads: processing capacity, availability, efficiency, reliability, completed jobs, latency, and bandwidth.
White-paper pipeline
Resource contribution
Resource measurement
Computational verification
Network recognition
Layered architecture
Complexity below. A unified fabric above.
02
Coordination layer
Work discovery, scheduling, routing, and dynamic balancing across heterogeneous nodes.
03
Verification layer
Result commitments, timing, redundant checks, signed receipts, and reputation derived from observed performance.
04
Application layer
A unified interface through which AI, rendering, simulation, and data workloads can consume capacity.
Nodes remain autonomous
Contribute on your own terms.
The architecture is hardware-agnostic. A node operator chooses utilization limits, power and thermal boundaries, schedules, and which resources are available. Coordination should route work around those constraints—not erase them.
Trust from evidence
Reputation follows performance.
Verification measures whether authorized work was completed accurately and on time. Signed results, repeatable checks, and operational history allow reliable nodes to build reputation without assumed authority.
Explore Proof of Resource →What the fabric can serve
One resource pool. Many kinds of work.
AI workloads
Training, inference, and distributed model operations.
Graphics
Rendering, spatial computing, and visual production pipelines.
Simulation
Parallel scientific, engineering, and physical-system workloads.
Data
Large-scale processing that can be divided into verifiable tasks.
Running pilot
The building blocks are observable.
GRID currently operates a public Genesis node, P2P replication, privacy-preserving node telemetry, a coordinator, verified settlement receipts, and Genesis-signed blocks.
Verify the live network →Vision boundary
The paper describes the destination.
A planetary permissionless compute marketplace, decentralized block finality, universal workload support, and mature token economics remain roadmap goals. They are not presented here as completed mainnet features.
Original document
Read the complete architecture and vision.
Download the original ten-page GRID white paper as a PDF.