Energy module · Live · Data current as of May 2026
Reasoning, when markets break.
A propagation engine for the global energy and supply-chain system.
1,200 assets across 86 countries and 19 commodity classes. Click any
one — the engine traces the cascade. Backtested against 71 historical
events with 96% accuracy within ±1 tier.
Atlas is optimized for desktop. Bookmark this page and revisit on a larger screen for the full experience.
The gap
Energy markets move at the speed of geopolitics.
News
Headlines tell you what just happened.
Dashboards
Track what's already happening.
Neither
Tells you what happens next.
The instrument
Map. Model. Simulate.
Three layers, separable by design. Engine reasons procedurally. Logic encodes
decisions. Data points are declared with confidence tiers. You can interrogate
any assertion at any depth.
01
Map
1,200 assets — refineries, terminals, pipelines, offshore and
onshore projects, power plants, mines, ports, chokepoints — across
19 commodity classes. Geographic and topological at once.
1,200 assets · 86 countries · 19 classes
02
Model
Dependency graphs with declared confidence. Every assertion is verified,
evidenced, reasoned, or estimated — visible at the point of use. The
platform is honest about what it does and doesn't know.
Four-tier confidence · auditable
03
Simulate
Arm any asset. The engine propagates the disruption through the graph
and returns one of six equilibrium archetypes. Calibrated against
71 historical events from 1999–2025.
Six archetypes · 96% backtest accuracy
The full methodology — including the per-event audit table, damper definitions,
amplifier weights, and the held-out backtest record — is published in detail.
Start here
Four scenarios. One click each.
The fastest way to see what the engine does is to watch it cascade through a
canonical disruption. Pick one to launch the Atlas pre-armed.
The platform reasons across 19 commodity classes — not just oil and gas,
but the full supply-chain web that ties them to fertilizer, freight,
industrial metals, and food.
● Live
Energy
Refineries · terminals · pipelines · projects · power
● Live
Metals & mining
Iron, copper, nickel, lithium, rare earths
● Live
Agriculture & food
Grain · palm oil · sugar · fertilizer
● Live
Logistics
Container ports · air freight · chokepoints
Calibration
Backtested. Not back-fitted.
The engine is exercised against five decades of real disruption events.
Where the model agrees with history, that's signal. Where it doesn't, the
delta is published.
1,200
Critical assets mapped
86
Countries covered
71
Historical events in calibration set
96
% backtest accuracy within ±1 tier
The position
We don't sell forecasts.
We build reasoning instruments.
Strata-intel is for the analyst, the trader, the policy desk, and the risk
committee that needs to see the cascade — not just the headline.
For
Strategy offices, sovereign wealth, policy desks, energy operators — anyone who needs structured answers to "what if X breaks", not dashboards.
What it's not
Not a price prediction engine, not a news terminal, not a consultant deck. The engine reasons about physical supply chains; pricing is downstream.
A propagation engine for the global energy and supply-chain system. 1,200 assets across 86 countries and 19 commodity classes. Click any one and the engine traces the cascade — who gets hit, by how much, on what horizon. Calibrated against 71 historical events; 96% backtest accuracy within ±1 tier.
For
Strategy offices, sovereign wealth, policy desks, energy operators — anyone who needs structured answers to “what if X breaks”, not dashboards.
What it’s not
Not a price prediction engine, not a news terminal, not a consultant deck. The engine reasons about physical supply chains; pricing is downstream.
73 countries scored on net energy position, import dependency, and chokepoint exposure. Click any country for full profile breakdown.
Sort by
Tier
#
Country
Region
Score
Independence (0–100)
Tier
Inspector
How the engine reasonsread-only
The platform's reasoning is grounded in three separable layers: engine (procedural code · how the system reasons), logic (model decisions · taxonomies, parameters, lens definitions), and data points (entity declarations · what facts the system knows). This view inspects the logic layer.
Fragility analysis
Probability of equilibrium breachPhase 3
Monte Carlo simulation across the parameter distributions declared in MODEL_CONFIG.fragility_params. Each archetype carries its own distribution shape (μ, σ) from observed historical disruption magnitudes. Output is the probability that the new equilibrium clears within tolerance — and what regime it lands in.
Calibration
Engine vs. realityPhase 4
The engine's reasoning is honest only if it can be checked. This view replays curated historical disruption events through the engine and scores predictions against actual observed outcomes. Each backtest yields a quantitative match score plus actionable calibration insights — over-predictions, under-predictions, and gaps where the engine missed reality.
Per-asset score
Asset Stress Index
A 0–100 score measuring expected damage to the global commodity system if an asset were fully disrupted for 30 days. Composed of throughput share, substitutability (inverted), dependent-country concentration (Herfindahl), and cross-commodity coupling. Calibrated against historical events.
Reports
Generate a report
Select one or more report types. Each is a PDF generated on demand from the engine. Pick a subject in each panel and click Generate.
Chokepoint
For any major transit chokepoint: who depends on it, what flows through, what happens if closed for 30 days, bypass options, multi-commodity cascade analysis, and historical incidents.
Country
Energy and commodity supply-chain vulnerability assessment for any of ~86 countries with curated dependency portfolios.
Commodity
For any commodity class: top producers, top importer countries, transit-chokepoint exposure, and substitution-ceiling statistics. Surfaces structural concentration.
Operator
For any major operator: full asset portfolio, geographic concentration, and worst-case single-asset failure scenarios. Surfaces single-point-of-failure exposure.
⚡
Run a disruption scenario
Click any asset → arm → see global cascade
Layers
Click any entity for details. For country-level analysis, see Country Scoring tab.
Pick layers to choose which assets to disrupt. Click any visible asset → ⊕ Arm scenario. Multi-select supported for compound disruption.
Energy Independence
85+ Major exporter
70-84 Self-sufficient
50-69 Partially dependent
30-49 Highly dependent
<30 Critically dependent
Larger circle = higher score
Pipelines
Oil
Gas
Refined products
HVDC
Dashed = planned/idled
Terminals + STS
LNG export (square)
LNG regas (faded circle)
Crude export (circle)
Product / bunker (diamond)
STS zone (dashed circle)
Sample: ~60 entities. Live atlas backend will populate to 150+ globally. Includes shadow-fleet STS hubs (Laconia, Lome, Vostochny) with sanctions context.
▾ Scenario Builder
Click any asset → ⊕ Arm scenario → set severity + duration → run. Multi-select supported.
Armed for compound
Substitution flow
Unmet demand
Every typed entity is armable. Curated entities use evidenced paths; others use defaults.
▾ SCENARIO BUILDER
Build any disruption you want to reason about
The platform's reasoning engine works on any asset you arm. Click an asset on the map, set severity and duration sliders, run. Multi-select supported — armed entities run as a compound scenario with congestion-aware substitution allocation and predicted price impact.
or browse pre-built examples
Or try a showcase scenario
Each runs the live engine. Reasoning track explains why every tier is what it is.
Picks two showcases and produces a side-by-side comparison PDF. Same engine, same gates, same reasoning — just two outputs diffed.
● click any asset to arm
● two sliders per entity
● multi-select for compound
● share via URL
Coming in v37+
Onshore Projects
Description.
Global Capex Timeline · 2026
Data Sources
Compare Projects
Scenario Library
Strata-intel · v192 · May 2026
Reasoning instrument for cross-commodity disruption scenarios.
What this is
A reasoning instrument for global commodity supply chain disruption. Click any asset on the map — refinery, port, mine, fertilizer plant, grain corridor, container terminal — and the engine computes how its disruption propagates across countries and across commodities. Multi-select supported — armed entities run as a compound scenario with cross-commodity substitution.
Four tabs, four jobs
Atlas — explore infrastructure across 15 commodity domains (energy, metals, food, logistics, aviation). Toggle layers via the panel top-left. Click any asset for details.
Simulate — build disruption scenarios. Click any asset on the map, set severity & duration, run the cascade. Multi-select supported — armed entities run as a compound scenario.
Asset Stress — per-asset 0–100 score for every entity measuring expected damage if it went offline for 30 days. Two views: Global ranks every asset; Per Country scopes to a country's own infrastructure plus the external nodes its consumption depends on.
Reports — generate full PDF reports for any country, chokepoint, commodity, or operator. The chokepoint report renders the engine's complete output: 30+ pages of multi-commodity cascade analysis.
First time? Try this
On the landing page, click any of the four pre-cooked scenarios (Strait of Hormuz closes / Druzhba severed / Jamnagar offline / Sabine Pass outage). The engine arms the entity, runs the cascade, and shows you:
• A severity classification with diagnosis (e.g. "Global Deficit / Price Spike")
• Primary flow disrupted, substitution capacity, gap (mbpd)
• Country-by-country impact tiered by severity (catastrophic / severe / manageable)
• A trajectory showing how the gap evolves week-1 -> month-1 -> month-6
• A "Download PDF" button on the result panel
Then go to Asset Stress and pick a country from the dropdown to see that country's stress-ranked exposure.
How accurate is it?
Accuracy varies by event type. Full backtest numbers, per-event audit, and known limitations are published on the methodology page.
This is a structured estimator for physical supply chain reasoning, not a price prediction engine. It tells you whether a disruption is moderate or severe; it does not tell you whether Brent goes to $95 or $105. It models supply-side fundamentals well; it does not model speculative positioning or refinery utilization games. Treat output as a first-cut estimate that an analyst with desk-specific knowledge would refine.
Pre-built scenarios in the library covering chokepoints, sanctions, palm, port closures
Class-aware severity classification with three independent tier signals
Winners alongside losers — every disruption identifies who captures the lost share
What's new in this version
v192 ships the engine-fidelity rebuild of the chokepoint stress report. Every field the engine produces is now rendered in the PDF — 12 distinct product classes per chokepoint (was 1), full per-country product lists in the country exposure tables, 3-horizon weight trajectories inline, a new entity-level cascade traversal page showing hop-by-hop graph propagation, aggregate tertiary-impact summaries for backfill chains, and engine diagnostics in the audit trail. The Hormuz report is now 30+ pages of multi-commodity cascade analysis instead of an oil-focused summary.
Roadmap
Country / commodity / operator PDF reports rebuilt to the same engine-fidelity standard as the chokepoint report. Expanded country portfolios with 5-12 products per dependency edge (currently 1-4). New entities for missing commodity classes — Persian Gulf aluminum (EGA, Alba, Sohar), Persian Gulf fertilizer (QAFCO, SAFCO, Fertil, GPIC, OMIFCO), helium (Qatar Ras Laffan + Algeria + US), sulfur, ammonia, phosphates. Live data integration (AIS, prices, weather). Backend extraction for the dependency graph. Mobile-optimized layout.
Feedback
What's working, what's not, what's missing
Direct feedback shapes the next version. Specific is better than general — "the arm button on Hormuz didn't update when I clicked it" beats "UI feels wrong." Bug reports, missing assets, calibration corrections, feature ideas — all welcome.
⚠ Disruption Simulator
Strata-intel is free during beta.
Drop your email and run scenarios. We'll email you when the Pro tier launches with unlimited scenarios, scenario stacking, and live-data refresh.
Run Disruption Scenario
Hormuz Closure Simulation
API Integrations
Global Fishing Watch
Disconnected
Free SAR-detected offshore platforms (oil/wind/unknown) from 2017–present.
Get a token at globalfishingwatch.org/our-apis (free, requires impact statement).
Platforms appear on the Rigs layer when zoomed in.
Open-Meteo
Connected · No Key Required
Free weather data with no API key needed and CORS enabled. Used for current conditions in project detail panel.
See open-meteo.com. Auto-fetched on project select.
NOAA NHC Active Storms
CORS-Restricted
NOAA publishes free GeoJSON via the NHC ArcGIS endpoint at mapservices.weather.noaa.gov/tropical.
Direct browser fetch is CORS-blocked — production deployments need a backend proxy.
Active storms shown on map are seed data; tap an Atlantic storm to see the real path NOAA would return.
Note: API keys are stored in browser memory only — they're cleared when you close the tab. For production, keys belong on a backend with a CORS-proxy layer.