# Quantum Motion adds prototype CUDA-Q backend to QUICHE chemistry framework

Published 2026-11-19 · Quantum.world Team · Quantum.world

> Quantum Motion added NVIDIA CUDA-Q as a prototype backend to its QUICHE chemistry framework. A demonstration runs a small iron-sulfur chemistry example from an ORCA calculation through quantum phase estimation and simulation on NVIDIA accelerated computing.

A new prototype backend lets QUICHE pass chemistry workflows into NVIDIA CUDA-Q for simulation and resource estimation, [Quantum Motion](https://quantummotion.com/a-fully-integrated-end-to-end-quantum-chemistry-pipeline-with-quiche-and-nvidia-cuda-q/) said in a post dated Nov. 19. The company’s demonstration starts with a calculation in the ORCA chemistry package, sets up a quantum algorithm through QUICHE and CUDA-Q, then runs a simulation on NVIDIA accelerated computing.

## Connecting the software

The work brings together tools with distinct roles in the workflow. ORCA runs the classical chemistry calculations and exports active-space integrals in the standard FCIDUMP file format. QUICHE reads that file and lets a user specify the Hamiltonian, the flavor of quantum phase estimation, the simulation method and a target energy error.

CUDA-Q Algorithms supplies high-level abstractions for QUICHE to use, including construction of qubit Hamiltonians. The quantum phase-estimation circuits generated through QUICHE run as CUDA-Q kernels on NVIDIA’s high-performance backend. Quantum Motion describes the user-facing workflow as a short Python notebook, with no hand-written circuits or manual conversion steps between the ORCA input and calculated energies.

The same QUICHE specification used to run the simulation also serves as the starting point for resource estimation. Quantum Motion says the pipeline stays virtually unchanged for resource estimation in fault-tolerant quantum computing.

## A small iron-sulfur example

The demonstration examines spin-state energetics in an iron-sulfur system, a class of chemical problems relevant to understanding reactivity, electrochemistry and photochemistry. The chosen molecule is Fe(SH)4−, modeled with an active space of five electrons in five orbitals.

Quantum Motion says the example was selected as a minimal model of the iron site in rubredoxin, which the post describes as the simplest natural iron-sulfur protein. It is deliberately small: the company says its size lets the team compare each quantum phase-estimation result with an exact reference from complete active space self-consistent field calculations, and run the full pipeline on a single GPU.

Spin-state energy differences matter in the example because states with different multiplicities can be close in energy. The post notes that some iron-sulfur clusters have reported differences within sub-milliHartree, and that different methods and choices about including dynamic correlation effects can produce different spin-ladder orderings.

## Prototype, not a hardware result

The new connection is a software prototype and the published demonstration is a simulation on NVIDIA accelerated computing, not a result from a quantum processor. Quantum Motion, FACCTs and [Riverlane](https://quantum.world/companies/riverlane) developed QUICHE as a project to bring quantum computing into standard quantum chemistry workflows, including workflows using ORCA.

Quantum Motion says the integrated pipeline is meant to support chemistry simulations and resource estimates, including for systems that are no longer classically simulable. The post links to the full demonstration.

## Sources

1. Original source [Quantum Motion](https://quantummotion.com/a-fully-integrated-end-to-end-quantum-chemistry-pipeline-with-quiche-and-nvidia-cuda-q/) quantummotion.com

## Related stories

- [NVIDIA releases QEC decoder, claims up to 13,000× speedup](https://quantum.world/news/nvidia-releases-tcbp-jobs-qec-decoder)
- [Quantum chemistry study tests two UCJ initialization methods](https://quantum.world/news/ucj-quantum-chemistry-initialization-methods)
- [Researchers train chemistry circuit generator on 70-qubit hardware](https://quantum.world/news/hardware-trained-70-qubit-chemistry-circuit-generator)
- [PNNL Breaks a 164-Qubit Chemistry Model into 12-Qubit Tasks](https://quantum.world/news/pnnl-qflow-164-qubit-chemistry-model)
- [SQD Recovers FeO Model Ground States from IBM Quantum Samples](https://quantum.world/news/sqd-recovers-feo-ground-states-ibm-quantum-samples)

## Explore

[More articles](https://quantum.world/news)

---

Canonical: https://quantum.world/news/quantum-motion-quiche-cuda-q-chemistry-pipeline
