Quantum PDE Solvers: An Honest Verdict
Quantum algorithms for differential equations exist on paper, but for Kronos plasma and field solves the caveats erase the advantage.
Why PDEs tempt and disappoint
Kronos runs heavy PDE workloads: free-boundary equilibrium (Grad-Shafranov), gyrokinetic turbulence, MHD stability, neutronics transport. Because quantum states live in exponentially large spaces, it is tempting to hope quantum PDE solvers offer exponential speedups. In practice they inherit the same limits as HHL: readout, state preparation, conditioning, and nonlinearity.
# Typical quantum PDE route: discretize -> linear system A x = b -> HHL-like
# Exponential state-space advantage is undone by:
# readout : need field values, not <x|M|x> scalars -> O(N) cost
# loading : encoding source/boundary data -> state prep cost
# nonlinear : Navier-Stokes / gyrokinetics not linear -> linearization
# kappa : stiff operators -> kappa^2 blowupThe nonlinearity wall
Quantum evolution is linear and unitary, but plasma transport and turbulence are strongly nonlinear. Encoding nonlinear dynamics requires Carleman-style linearization (embedding into a larger linear system) that is only accurate at weak nonlinearity and grows the dimension, or repeated measurement-and-reprepare loops that shed any speedup.
What Kronos actually does
- Equilibrium and MHD: classical FEM/spectral solvers and physics-informed neural networks, validated in the twin.
- Turbulence: classical gyrokinetic codes on GPU/HPC, with neural surrogates for fast lookups.
- No quantum PDE solver is on the operational or near-term research critical path.
This page exists to close a door cleanly. Claiming quantum speedups for plasma PDEs would be exactly the kind of overclaim we refuse. The genuine long-term quantum opportunities for Kronos are in materials Hamiltonian simulation and, more speculatively, amplitude-estimation Monte Carlo, not field PDE solves. The roadmap reflects that ordering.