Quantum Systemic Oracle - Crypto Risk
Daily crypto systemic-risk score - IonQ Forte Hardware

QAOA Recommended Portfolio - the lowest-risk set of assets to hold that the quantum optimizer (QAOA running on IonQ) actually found and sampled. This is the recommendation the oracle publishes. It can hold a slightly different number of assets than the classical optimum below (for instance four names instead of three). That is expected, not an error: this is a budget-dominated problem where many baskets sit within a few percent of each other on cost (that spread is the gap shown up top), so the quantum run settles on a near-optimal neighbor of the brute-force best, sometimes carrying one extra, nearly cost-neutral asset, rather than reproducing the exact same set.

Classical Optimum (brute-force) - the provably best hold set, found by brute force: exhaustively scoring every one of the 2^N possible hold/skip combinations (2⁸ = 256 for this 8-asset cohort) and keeping the lowest-cost one. It is the ground-truth benchmark, but brute force only stays feasible because the cohort is small.

How QAOA compares (the "gap") - for a small cohort the brute-force optimum is unbeatable on cost, so QAOA's job is to match it. The QAOA gap vs optimum (shown above) measures how close it came; a near-zero gap means the quantum run recovered optimal-quality answers. QAOA's real advantage is scale: brute force explodes as 2^N (≈10¹⁸ combinations at just 60 assets*) and quickly becomes impossible, while QAOA stays tractable - so it is the path to large cohorts a classical search could never enumerate.

*What 60 assets needs on hardware - this encoding uses only N + 6 qubits (N asset-decision + 6 global-witness), so a 60-asset book is ≈66 logical qubits, not 60. That is within reach of IonQ's upcoming Tempo (100 physical qubits, #AQ 64, 99.9% two-qubit fidelity). Crucially, the binding limit here is two-qubit-gate error, not qubit count: Tempo's ≈10× lower gate error vs Forte widens the usable two-qubit-gate envelope by roughly the same factor (Forte's today is ≈90 gates), so the deeper, fully-coupled circuit a ~60-asset QAOA requires would stay structured where Forte flattens toward uniform. Fittingly, #AQ 64 (≈2⁶⁴ usable states) lands right at the ≈10¹⁸-combination scale this 60-asset problem reaches - so a ~66-qubit Tempo run is the natural home for a cohort this size. At its full 100-qubit width the same N + 6 encoding tops out around ≈94 assets (100 − 6 witnesses) - a cohort whose ≈2⁹⁴ (≈10²⁸) combinations are hopelessly beyond any classical brute force. (Order-of-magnitude estimate; actual capacity depends on circuit depth, coupling, and error mitigation.)