NOURIX — SYSTEM BLUEPRINT

Nourix

Open Compute Infrastructure for AI

Connect workloads with decentralized computational resources through an open marketplace for AI inference, agents, simulations, rendering, data processing, and more.

Explore Network
SEC.02THE COMPUTE NETWORK

Feed Intelligence With Open Compute

FIG. 2.1 — NETWORK TOPOLOGYWORKLOAD CREATOR01NOURIX ROUTING02PROVIDER MARKET03EXECUTION04VERIFICATION05SETTLEMENT06
SEC.03COMPUTE MARKETPLACE

Compute, Priced by the Network

DEMO / INTERFACE DATA — NOT LIVE NETWORK DATA
ProviderHardwareAvail.Perf.Rep.Price/HRStatus
MERIDIAN LABSPX-0644CPU Cluster97%624.3$0.24ONLINE
CIRRUS NODEPX-0915RTX 409095%824.4$0.58ONLINE
NORTHGRIDPX-0402RTX 409099%844.6$0.61ONLINE
KAPPA SYSTEMSPX-0518L40S92%874.5$1.18BUSY
DELTA COREPX-0809NVIDIA A10074%894.2$1.78OFFLINE
VECTOR FORGEPX-0233NVIDIA A10088%914.7$1.92ONLINE
ORION RACKPX-0781NVIDIA H10081%964.8$2.71ONLINE
HELIOS COMPUTEPX-0117NVIDIA H10096%984.9$2.84ONLINE
SEC.04WORKLOAD SUBMISSION

Deploy Workloads Across Open Infrastructure

FORM 4.1 — WORKLOAD SPECIFICATION

FRONTEND / DEMO BEHAVIOR. NO PROTOCOL CONTRACTS ARE CONNECTED — NO JOBS ARE BROADCAST.

FIG. 4.2 — ROUTING SEQUENCE
  1. JOB CREATED
  2. SEARCHING PROVIDERS
  3. MATCHING
  4. PROVIDER SELECTED
  5. EXECUTION SESSION CREATED

SUBMIT A WORKLOAD TO RUN THE ROUTING SEQUENCE.

SEC.05PROVIDER NETWORK

Supply Compute. Power Intelligence.

Infrastructure providers make computational resources available to the Nourix marketplace — GPUs, CPUs, servers, and storage. Idle capacity becomes productive network infrastructure, and providers earn by executing workloads.

FIG. 5.1 — RACK ELEVATION / ROTATABLEHOVER COMPONENTS TO INSPECT
COMPONENT INSPECTOR

GPU 0 — H100

GPU
H100 SXM
VRAM
80 GB
CAPACITY
989 TFLOPS
AVAILABILITY
96%
REPUTATION
4.9
JOBS
12,480
PROVIDER ONBOARDING

Independent operators can contribute available resources and compete for workloads. No affiliation with a single organization is required to participate.

SEC.06ROUTING ENGINE

Workloads Routed by the Network

FIG. 6.1 — MATCHING SCHEMATICPRICEPERFORMANCEAVAILABILITYREPUTATIONREQUIREMENTSNOURIXROUTING ENGINEOPTIMALPROVIDEREVALUATION: COST / PERF / AVAILABILITY / HISTORY / CAPACITYOUTPUT: DYNAMIC ALLOCATION TOWARD HIGHEST-DEMAND RESOURCES
COST

Providers compete on price, driving compute toward market rates instead of fixed list prices.

PERFORMANCE

Historical throughput and latency inform which infrastructure can meet execution targets.

AVAILABILITY

Only infrastructure that is actually online and capable receives demand.

REPUTATION

Track records of reliable completion improve a provider’s access to future workloads.

REQUIREMENTS

Workload-specific constraints — resources, latency, verification — constrain the match set.

SEC.07EXECUTION & VERIFICATION

Compute You Can Verify

FIG. 7.1 — JOB LIFECYCLE
  1. 01JOB SUBMITTED
  2. 02PROVIDER MATCHED
  3. 03EXECUTION SESSION
  4. 04RESULT SUBMITTED
  5. 05VERIFICATION
  6. 06SETTLEMENT
DETERMINISTIC VERIFICATION

Results can be checked against known computational rules where possible.

REDUNDANT EXECUTION

Certain workloads can be executed independently to compare results.

CHALLENGE MECHANISMS

Network participants can challenge questionable execution or submitted results.

PROVIDER COLLATERAL

Providers can commit collateral against certain workloads, creating an economic incentive for honest execution.

Verification requirements vary with the characteristics and risk profile of each workload. No single mechanism provides complete security on its own — collateral, reputation, verification, challenges, redundant computation, transparent settlement, and onchain records work together as layered protection.

SEC.08MULTI-WORKLOAD ARCHITECTURE

One Marketplace, Every Workload

AI INFERENCE

Run model inference through decentralized compute providers.

AGENT EXECUTION

Provide computational environments for autonomous AI agents.

MODEL EVALUATION

Benchmark and evaluate models across distributed infrastructure.

DATA PROCESSING

Execute compute-intensive data pipelines.

SIMULATION

Access computational resources for scientific and economic simulations.

RENDERING

Distribute rendering workloads across available providers.

CUSTOM COMPUTE

Support additional computational workloads through the open marketplace.

+ HIGH-PERFORMANCE WORKLOADS
SEC.09NETWORK ECONOMICS

Compute Becomes an Open Market

FIG. 9.1 — ECONOMIC FLYWHEELWORKLOAD USERSPROVIDERSSETTLEMENTVERIFICATIONNOURIXASSET
WORKLOAD USERSPAYMENTSPROVIDERS
PROVIDERSCOLLATERALNETWORK
NETWORKINCENTIVESPARTICIPANTS
RESULTSVERIFICATIONSETTLEMENT

The native Nourix asset is intended to coordinate economic activity across the network — workload payments, provider settlement, provider collateral, network incentives, verification participation, and marketplace coordination. By connecting economic activity directly to infrastructure demand, Nourix aligns network growth with actual computational usage.

NO TOKEN SUPPLY, ALLOCATION, APY, REVENUE SHARE, BURN RATE, PRICE, OR STAKING RETURN FIGURES ARE PUBLISHED. THE ASSET DESIGN FOLLOWS THE WHITEPAPER ONLY.

SEC.10NETWORK ARCHITECTURE

The Stack, Drawn Like a Drawing

FIG. 10.1 — SYSTEM ARCHITECTURE / SECTION VIEW
L4APPLICATION LAYER
  • AI Applications
  • Agents
  • Developers
  • Data Pipelines
L3NOURIX MARKETPLACE
  • Job Submission
  • Provider Discovery
  • Pricing
  • Routing
L2EXECUTION LAYER
  • Compute Providers
  • GPU Clusters
  • CPU Infrastructure
L1VERIFICATION LAYER
  • Execution Validation
  • Provider Reputation
  • Job Results
L0SETTLEMENT LAYER
  • Payments
  • Collateral
  • Incentives
  • Onchain Settlement
SCALE: LOGICALSHEET 10 OF 13 — REV A
SEC.11GLOBAL TOPOLOGY

A Distributed Layer of Providers

FIG. 11.1 — ABSTRACT NETWORK TOPOLOGYNETWORK VISUALIZATION / DEMO
NORTH AMERICASOUTH AMERICAEUROPEASIA PACIFICFICTIONAL DEMO NODES — NOT REAL PROVIDER LOCATIONS
SEC.12TECHNICAL BLUEPRINT

Nourix Technical Blueprint

DOC NX-WP-001REV A

Nourix

INFRASTRUCTURE THAT FEEDS INTELLIGENCE

Nourix is an open infrastructure marketplace designed to connect AI applications and computational workloads with independent infrastructure providers.

SHEET 1 / 22
KEY CONCEPTS
  • 01The Nourix Network
  • 02Workload Marketplace
  • 03Intelligent Routing
  • 04Execution Sessions
  • 05Verification
  • 06Provider Reputation
  • 07Economic Layer
  • 08Security Model
  • 09Settlement
  • 10The Network Flywheel
Read Whitepaper
SEC.13DOCUMENTATION

Build With Nourix

GITBOOK / NX-DOCS

Explore the architecture, network concepts, compute marketplace, provider model, workload lifecycle, verification, settlement, and integration concepts behind Nourix.

Open GitBook