High-performance computing for financial modelling and wargaming

High-performance computing for financial modelling and wargaming is the use of substantial computing capacity, data and models to test financial and economic scenarios. It is analytical infrastructure, not a coercive instrument. Public evidence establishes extensive supervisory stress testing and financial-stability modelling. The sources establish modelling workflows and computationally demanding methods, not a common hardware threshold for what counts as high-performance computing. They do not establish that every government uses the same systems for offensive economic-warfare planning.

Function

Supervisory stress tests project losses, income and capital under hypothetical conditions. The Federal Reserve's 2026 exercise includes a macroeconomic scenario, a global market shock for firms with large trading operations and a largest-counterparty default component for specified firms. The European Banking Authority's 2025 exercise tested banks under a severe hypothetical macro-financial deterioration. These are resilience exercises, not forecasts.

Agent-based models address a different problem. They simulate heterogeneous institutions and feedback among funding, liquidity and asset prices. The United States Office of Financial Research and Bank of England researchers have used the approach to study crisis propagation. Such models can inform sanctions, blockade or supply-shock scenarios, but an analytical possibility is not proof of operational adoption.

Strategic significance

Strategic use requires secure data, validated assumptions and a defined decision process as well as compute. Possible applications include dependency mapping, scenario comparison, spillover analysis, evasion testing and post-action assessment. More computing power does not remove uncertainty or establish causation.

Control or weaponisation history

Models are vulnerable to missing data, structural breaks, strategic adaptation and reflexive behaviour. Results depend on scenario design and validation. Assessment should record model purpose, input date, sensitivity tests, uncertainty and the decisions actually informed. Simulation can discipline judgement; it cannot make coercive outcomes reliably predictable.

See also

Economic Kill Chain (EKC) · Compute as strategic resource · Reflexive control · Taiwan blockade scenario (economic dimension) · Financial intelligence (FININT) · Algorithmic and high-frequency trading systems

Sources

1. Federal Reserve Board, *2026 Stress Test Scenarios* (2026).
2. European Banking Authority, *2025 EU-wide Stress Test: Results* (2025).
3. Office of Financial Research, "An Agent-based Model for Crisis Liquidity Dynamics" (2015).
4. Bank of England, "Agent-based modelling at central banks: recent developments and new challenges" (2025).