Infrastructure.
Power, water, ports, telecom, and rail — where one node is the system.
Infrastructure cascades follow physics, not procurement. A single substation, a single subsea cable landing, a single rail interchange can take a region offline — and the dependency map that would surface those nodes in advance often only exists in operations databases, not in the models that price the risk.
Standard contingency analysis (N-1, N-2) is a regulatory floor, not a structural map. It enumerates failure cases but not the cascade depth — which other infrastructures, which counterparties, which downstream production sites fail with the named one. The relevant graph crosses the seams: power into telecom into finance into manufacturing.
When something does fail, the loss isn't bounded by the asset. It's bounded by everything the asset was carrying. That's the number a CFO, a reinsurance underwriter, or a national-security planner needs, and it's the number nobody computes ahead of time.
Manifold ingests the physical-asset graph (substations, lines, transformer banks, cable landings, water mains, pipelines, rail interchanges) and joins it to the dependency layer above (telecom carriers, settlement systems, hospitals, manufacturing plants, defense logistics). Edges encode physical, contractual, or jurisdictional dependency — same schema, different sources.
PEARL runs counterfactuals on named outage scenarios (this substation goes dark for 72 hours; this cable landing is severed). PARETO simulates the resulting cascade across the joined graph. SPIRTES discovers structural dependencies that operations data implies but no formal map captures.
Regulatory N-1 contingency analysis ignores the real cascade graph. The substation, transformer bank, or HV interconnection that decides regional reliability is buried in operating data, not surfaced as a structural risk.
ΩF score per node on the live grid graph. Counterfactual outage simulation across substations, lines, and interconnections. Decisive-node ranking under contingency stacks.
Property-cat models price physical asset damage. They miss the cascade — the cable landing that takes telecom and finance offline, the port that backstops half the regional supply chain.
Network-aware exposure across the insured graph. Cascade load and tail-depth on the lines you actually wrote.
Strategic-asset lists rank by size or visibility. The decisive nodes — the ones whose failure cascades — aren't the same set, and aren't ranked anywhere.
I + C ranking on the national infrastructure graph, configurable per strategic objective (continuity of government, defense logistics, financial settlement).
Run Manifold on a infrastructure graph.
Trial accounts come pre-loaded with a curated dataset. Or request an invite to bring your own graph.
