The Global Risks Index · Kernel · 15 products · 35 domains
One kernel for risk intelligence
The kernel resolves context, finds the right governed science, runs it and composes products that keep every input’s evidence, supported claim level and correction path.
Page GRX-KRNEdition Baseline Edition 2026Evidence close 2026-09-30Values not yet publishedMethod v1.0.0-draft
15
kernel products
35
frontier domains
3
integration profiles
50
product kinds
01Kernel
What the kernel does
The risk engine sits over the Foundry systems and does not duplicate NXS-GRX (canonical data, provenance and risk) risk science; it composes what the science already produced.
Context
Resolve
Places, times, assets and portfolios resolve to UNIs, with every source assertion kept.
Plan
Discover and route
Capabilities are found by stable ID and version, never by title, and routed by latency and scientific assurance.
Run
Run and compose
A dependency plan runs on governed NXS-GRX kernels; products compose across domains on the risk product graph.
Deliver
Project and correct
Products project to the form each client needs, and a correction upstream supersedes every dependent product.
02A composition
From request to product
Every composition is a governed run: the policy gate refuses with a type, the assurance gate fails closed, and the product carries its lineage.
on: product requested
Compose a risk intelligence product with the risk engine
In progressTriggered by You or your system10 jobs1 approval gate1 matrix job5 checks1 deployment
Graph
01Context
02Context
03Plan
04Plan
05Plan
06Run
07Run
08Check
09Deliver
10Deliver
100%
Timeline
Job0 h1 hTarget
Resolve contextRisk engineinstant
Purpose and rights checkIdentity and policy authorityinstant
Discover capabilitiesRisk engineinstant
Route by quality of intelligenceRisk engineinstant
Compile the DAG planRisk engineinstant
Run governed scienceNXS-GRX run serviceoperational
Compose across domainsRisk engineinstant
Evidence and assurance gateRisk engineinstant
Project the productRisk engineinstant
Lineage and correctionRisk enginecontinuous
End to end on the longest path: about 1.1 h of proposed target service levels, before the jobs that run with no fixed end. Gates wait for the people named; bars with a fading end have no fixed length.
Steps
-
Resolve contextContext · Risk engine · instantPlace, time, asset or portfolio resolved to UNIs · Purpose and tenant scope bound to the request · Source assertions kept, never merged silentlyRisk engineNS-1 UNI catalogue
-
Purpose and rights check approvalContext · Identity and policy authority · automaticIdentity and purpose verified · Rights, sovereignty and disclosure policy applied · Refusal is typed, never silentFoundryLicence register
-
Discover capabilitiesPlan · Risk engine · instantCapabilities found by stable ID and version, never by title · Implementation and release state read per capability · Held capabilities excludedRisk engineTested model kernels
-
Route by quality of intelligencePlan · Risk engine · instantPaths scored on latency class and scientific assurance · Fastest path that meets the assurance asked for · Trade-off recordedRisk engine
-
Compile the DAG planPlan · Risk engine · instantDependencies ordered topologically · Cycles refused · Plan versioned and hashedRisk engineNXS-GRX run service
-
Run governed science matrix ·Run · NXS-GRX run service · operationalEach node runs on its bound kernel · Seeds, inputs and versions recorded · Failures stay failuresNXS-GRX run serviceTested model kernelsMonte Carlo engine
-
Compose across domainsRun · Risk engine · instantProducts joined on the risk product graph · Weakest assurance carried forward · Every upstream evidence reference keptRisk engineProvenance (W3C PROV)
-
Evidence and assurance gateCheck · Risk engine · automaticSupported claim level never raised above its inputs · Missing evidence fails closed as
INSUFFICIENT_EVIDENCE· High-reliance projections need complete evidenceVerification engineProvenance (W3C PROV) -
Project the productDeliver · Risk engine · instantScalar, table, series, distribution, features, network or evidence pack · Agent-context, feature-vector and compact forms · Each projection carries its UNI and lineageRisk enginePublic cut APIVector tiles
-
Lineage and correctionDeliver · Risk engine · continuousDerivation and composition edges recorded · A superseded input marks every dependent product SUPERSEDED · Stale intelligence never stays currentCorrection logPostgreSQL ledger (row-level security)
Artefacts
| Artefact | Job | Who | Stage |
|---|---|---|---|
| Resolved context | Resolve context | Risk engine | Context |
| Policy decision | Purpose and rights check | Identity and policy authority | Context |
| Capability set | Discover capabilities | Risk engine | Plan |
| Route decision | Route by quality of intelligence | Risk engine | Plan |
| Execution plan | Compile the DAG plan | Risk engine | Plan |
| Upstream products | Run governed science | NXS-GRX run service | Run |
| Composite product | Compose across domains | Risk engine | Run |
| Assurance record | Evidence and assurance gate | Risk engine | Check |
| Projection | Project the product | Risk engine | Deliver |
| Lineage graph | Lineage and correction | Risk engine | Deliver |
| Correction notices | Lineage and correction | Risk engine | Deliver |
DoneIn progressWaiting for approvalQueuedApproval gateMatrix jobTest jobDeploymentYou are here
03Product truth
The kernel never upgrades science
- Supported claim level
- A composite is bounded by the supported claim level of its inputs. An integrated engine output stays integrated; it never becomes institutional acceptance.
- Assurance
- Composition carries the weakest scientific assurance of its inputs and every upstream evidence reference.
- Missing evidence
- High-reliance projections fail closed as
INSUFFICIENT_EVIDENCEinstead of inventing a value. - Ulungiso
- Superseding an upstream product marks every dependent composition SUPERSEDED, so stale intelligence cannot stay current.
- Non-claims
- The kernel issues no official warning, certifies no engineering, prescribes no treatment, executes no financial transaction and actuates no external system.
04Authority
Who decides what
| Authority | Owns |
|---|---|
| NXS-GRX | Canonical governed risk data, scientific products and evidence. |
| Identity and policy authority | Identity, purpose, rights, sovereignty, disclosure and claim boundaries. |
| Compute authority (NXS-COR (compute, nodes and runtimes)) | Execution placement, runtime and infrastructure lifecycle. |
| Capability lifecycle authority | Software, model and agent capability lifecycle. |
| Operational map authority | Operational map and common operating picture composition. |
| Risk engine | Composes, routes and projects governed risk intelligence. It does not manufacture scientific truth. |
| External institutions | Keep legal, regulatory, emergency, warning, medical, financial and other consequential authority. |
05Products
15 kernel products
3 of 15 products are release-held until their science and assurance evidence is complete. Products project to 50 kinds, from scalars and tables to rasters, meshes, networks, digital-twin deltas and evidence packs.
| Product | Into eyenzayo | Product kinds | Imeko |
|---|---|---|---|
| the risk engineunrisk.command | Operational entry surface for kernel status, model registry coverage, active compositions, products, corrections and integration posture. | status, capability graph, risk product graph | Integrated |
| NXS-STU (composition, dashboards and workflows) Composerunrisk.nexus-composer | Build arbitrary domain × lens × scenario × Chronotope compositions and compile them into evidence-bound scientific execution graphs. | time series, data cube, dependency graph, scenario tree, ensemble | Integrated |
| Chronotope Studiounrisk.chronotope-studio | Construct and inspect ordered observation, valid, forecast, scenario and transaction-time frames for cross-engine Nexus computation. | time series, trajectory, scenario tree | Integrated |
| model registryunrisk.risk-master | Search the complete NXS-GRX model registry, inspect stable capability IDs, methods, evidence, VVUQ obligations, implementation state and execution availability. | capability graph, model diagnostic, evidence pack | Release held |
| Scientific Runsunrisk.scientific-runs | Provider-neutral orchestration of fast kernels and durable Julia, Python, R and HPC scientific runs through canonical NXS-GRX execution authorities. | model diagnostic, ensemble, Monte Carlo sample, evidence pack | Release held |
| Risk Product Graphunrisk.product-graph | Inspect polymorphic risk products, derivation edges, provenance, evidence, uncertainty, corrections and supersession dependencies. | scalar, vector, matrix, tensor, table, time series, distribution, raster, coverage, data cube, mesh, network graph, trajectory, 3d scene, evidence pack | Integrated |
| Assuranceunrisk.assurance | Inspect numerical verification, calibration, validation, applicability, uncertainty decomposition, benchmarks and claim ceilings without promoting software state into scientific truth. | model diagnostic, benchmark result, evidence pack | Release held |
| Builder APIunrisk.developer | Discovery, API contracts, typed SDK metadata, templates, protocol profiles and semantic-conformance tooling for building any risk-aware application. | SDK manifest, API manifest, protocol profile, template | Integrated |
| Integration Studiounrisk.integration-studio | Discover and validate risk-intelligence projections for web, mobile, cloud, GIS, OT, IoT, XR, robotics, vehicles, digital twins, analytics and future platforms. | integration profile, conformance report | Integrated |
| Agentic Riskunrisk.agentic | Chronotope and evidence-bearing agent collaboration for querying, challenging, explaining and composing risk intelligence without granting agents truth or actuation authority. | risk agent context, evidence pack, time series | Integrated |
| Spatial Riskunrisk.spatial | Risk projections for Maps, GIS, digital twins, routes, fields, volumes, trajectories and spatial-computing clients while preserving actual spatial support and geometry provenance. | raster, coverage, vector features, vector tiles, mesh, point cloud, trajectory, route, 3d scene, digital twin delta | Integrated |
| Industrial Riskunrisk.industrial | Read-only risk-intelligence projection for critical infrastructure, industrial control, asset management and operational technology environments. | risk object, time series, dependency graph, state transition | Integrated |
| Edge Runtimeunrisk.edge | System-Pack-driven risk-intelligence deployment for regional, national, institutional, site, vehicle and device edge with explicit staleness and degraded-state semantics. | system pack, risk product cache, correction feed | Integrated |
| Risk Packsunrisk.packs | Versioned semantic, science, integration, policy and assurance compositions for technologies, disciplines, industries and deployment profiles. | domain pack, science pack, integration pack, policy pack, assurance pack | Integrated |
| Operationsunrisk.operations | Operational truth for plans, scientific runs, composition coverage, event subscriptions, corrections, materiality and safe scientific observability. | execution record, correction record, observability summary, subscription | Integrated |
06Frontier domains
35 frontier science and technology domains
The kernel’s domain registry covers the frontier technologies whose risks the index tracks, so a capability, product or pack always names its domain.
| Domain | Subdomains | Converges with |
|---|---|---|
| Artificial intelligence, machine learning and agentic systems | foundation models, multimodal AI, scientific ML, agentic AI, reinforcement learning, AI safety, AI assurance | – |
| Advanced computing and computational infrastructure | HPC, GPU accelerators, edge computing, cloud, neuromorphic, photonic computing, confidential computing | – |
| Semiconductors, microelectronics and chip systems | advanced nodes, chiplets, 3D integration, power electronics, RF, MEMS, advanced packaging, EDA | – |
| Quantum technologies | quantum computing, sensing, networking, communications, metrology, materials | – |
| Advanced communications and networked systems | 5G-Advanced, 6G, NTN, AI-RAN, O-RAN, MEC, optical networks, PNT | – |
| Cybersecurity, cryptography and privacy engineering | zero trust, post-quantum cryptography, confidential computing, OT security, identity, privacy engineering | – |
| Data and distributed digital infrastructure | data fabrics, knowledge graphs, distributed ledgers, DePIN, streaming, federated data | – |
| Robotics, autonomy and physical AI | industrial robotics, humanoids, UAS, autonomous systems, swarm robotics, field robotics | – |
| XR, spatial computing, human-machine interfaces and neurotechnology | AR, VR, XR, smart glasses, BCI, haptics, spatial interfaces | – |
| Biotechnology, genomics and synthetic biology | genomics, gene editing, synthetic biology, cell engineering, biofoundries, computational biology | – |
| Medical technology and computational health | digital health, medical devices, computational medicine, diagnostics, public health intelligence | – |
| Advanced materials and nanotechnology | nanomaterials, metamaterials, composites, 2D materials, smart materials | – |
| Advanced manufacturing | additive manufacturing, digital manufacturing, industrial automation, precision manufacturing, bio-manufacturing | – |
| Advanced energy systems | fusion, fission, renewables, storage, hydrogen, grid technologies, carbon management | – |
| Climate and carbon technologies | climate intervention, carbon removal, MRV, adaptation technology, climate modelling | – |
| Earth observation and geospatial intelligence | remote sensing, SAR, hyperspectral, geodesy, GIS, digital Earth | – |
| Space systems | satellites, constellations, launch, in-space servicing, space situational awareness, space communications | – |
| Aerospace and propulsion | advanced air mobility, hypersonics, electric propulsion, rocket propulsion, uncrewed aviation | – |
| Future mobility | connected vehicles, autonomous vehicles, rail automation, maritime autonomy, micromobility | – |
| Chemical and process engineering | process intensification, electrochemistry, catalysis, industrial chemistry, process safety | – |
| Civil and critical-infrastructure engineering | structures, bridges, dams, utilities, lifelines, resilience engineering | – |
| Water, hydrology, ocean and coastal science | hydrology, hydraulics, groundwater, oceanography, coastal engineering, water treatment | – |
| Agriculture, food and fisheries technologies | precision agriculture, crop science, aquaculture, fisheries, food systems, agri-robotics | – |
| Environmental and biodiversity science | ecology, conservation technology, biodiversity monitoring, ecosystem modelling | – |
| Geoscience, geotechnical systems and critical minerals | geophysics, geotechnics, mining, critical minerals, subsurface systems | – |
| Nuclear and radiological technologies | fission, radiological systems, fuel cycle, waste, radiation sensing | – |
| Photonics, optics and laser systems | photonics, lasers, lidar, optical sensing, integrated photonics | – |
| Frontier physics and plasma science | plasma, high-energy physics, condensed matter, accelerators, precision physics | – |
| Mathematics, statistics, operations research and complex systems | probability, optimisation, network science, dynamical systems, stochastic processes | – |
| Computational science, simulation and uncertainty quantification | numerical methods, multiphysics, digital twins, UQ, scientific computing | – |
| Metrology, sensing and autonomous scientific instrumentation | metrology, sensor networks, autonomous labs, scientific instruments, calibration | – |
| Industrial systems, reliability and process safety | reliability, FMEA, HAZOP, LOPA, SIL, asset integrity, OT | – |
| Financial engineering and computational risk | market risk, credit risk, insurance, actuarial science, systemic finance, quantitative finance | – |
| Supply-chain and industrial ecosystems | trade, logistics, supplier networks, critical dependencies, industrial policy | – |
| Convergent technologies and system-of-systems risk | cross-domain risk, emergent risk, technology convergence, systemic interactions, cascades | – |
07Integrations
Integration profiles
Each profile carries independent truth states: implemented is not tested, tested is not observed in operation, and none of these is a release.
| Profile | Umgangatho | Formats and transport | Truth state |
|---|---|---|---|
| unrisk.openapi.v1 | OpenAPI 3.2 | application/json, application/geo+json, HTTP/1.1, HTTP/2, HTTP/3 |
implementation: implemented interoperability: untested operational: unobserved release: hold claim: integrated |
| unrisk.mqtt.v1 | MQTT 5.0 | application/json, application/CBOR, MQTT/TLS |
implementation: registered interoperability: untested operational: unobserved release: hold claim: integrated |
| unrisk.agent.mcp.v1 | MCP 2026-07-28 | application/json, HTTP |
implementation: implemented interoperability: untested operational: unobserved release: hold claim: integrated |
08API
Where it runs
- Routes
GET /statusGET /frontier-stem/domainsGET /capabilitiesPOST /resolvePOST /discoverPOST /plansPOST /plans/:id/executeGET /products/:id/lineagePOST /products/:id/projectPOST /products/:id/correctGET /integrations- Access
- Tenant-scoped, inside authenticated NXS-GRX authority. Enterprise and public-sector plans reach it through the enterprise platform.
Compose with the risk engine
Enterprise and public-sector plans reach the kernel through the enterprise API.
Imeko
The methodology, structure and definitions of The Global Risks Index are a preparatory working draft (2026), circulated for review and comment and not for implementation. No value appears until an edition is released from a governed evidence cut; values then carry intervals, classes, sources and their correction pathway. Simulators run the published methods on numbers you choose and never show a NXS-GRX value.
What NXS-GRX is not
NXS-GRX is research and risk-analytics information. It is not a credit or ESG rating and uses no rating scale, not a benchmark for financial instruments, not official statistics, not an emergency warning or official classification, and not advice. It must not be used to evaluate natural persons. Risk classes are bands on a published scale, never rankings of countries or people.
Institutions and sources
The Nexus Consortium is stewarded by three independent not-for-profit institutions: The Global Centre for Risk and Innovation (GCRI), through GCRI US in Washington, DC and GCRI Canada in Ontario, with GCRI Canada as Lead Institution, Consortium Secretariat and lead contracting organisation; The Global Risks Forum (GRF); and The Global Risks Alliance (GRA). Technology, infrastructure and service providers work with the Consortium at arm’s length under written agreements, like every other external party in the public-private-planet model, and hold no role in its stewardship.
Every value carries the attribution of its sources, which keep their own licences. External names are cited for interoperability only. Nexus-authored text is intended for CC BY 4.0 on release.
Avis en français
La méthodologie et les définitions du NXS-GRX sont un projet de travail préparatoire (2026). Le NXS-GRX est une information de recherche et d’analyse des risques : ni notation, ni indice de référence financier, ni statistique officielle, ni alerte, ni conseil. Les classes de risque sont des bandes sur une échelle publiée, non des classements.