Ordoñez, Korpas & Kerenidis (Moody’s, HSBC, QCWare): Quantum Monte Carlo’s quadratic speedup, tested against reality

Key takeaways

• Quantum Monte Carlo promises a quadratic speedup over classical Monte Carlo by using amplitude estimation instead of repeated random sampling, which matters for pricing derivatives and running risk simulations that today take banks hours or days.
• The theoretical speedup only shows up once the number of qubits and circuit depth get large enough, so on current noisy hardware the quantum version can actually run slower than classical Monte Carlo for the problem sizes banks care about.
• Encoding real financial data and payoff functions into quantum circuits is itself a hard engineering problem, and the cost of that data loading step can eat into or erase the algorithmic advantage if it is not handled carefully.
• The group sees production use as still several years out, gated on error correction and better qubit counts, and argues the near term value is in building the software stack and benchmarking methodology now so teams are ready when hardware catches up.