Quanscient and Haiqu conducted a 15-step nonlinear simulation of quantum fluids.

      A new quantum algorithm has successfully executed a 15-step nonlinear fluid simulation around a solid obstacle on real quantum hardware, marking the most physically complex publicly recorded demonstration of such simulations. This technique decreases the needs for qubits and circuit depth, making industrial computational fluid dynamics (CFD) applications increasingly attainable.

      Quanscient, a Finnish simulation company, and Haiqu, a developer of quantum middleware, have showcased what they label the most intricate quantum computational fluid dynamics simulation conducted on actual hardware to date. The two organizations performed a 15-step nonlinear fluid simulation involving fluid dynamics around a solid shape—an issue pertinent to the design of aircraft wings and vehicle aerodynamics—on IBM’s Heron R3 quantum computer, utilizing a collaborative algorithm they created known as the One-Step Simplified Lattice Boltzmann Method (OSSLBM).

      CFD is a highly demanding field in engineering simulation, as it involves modeling fluid interactions with complex shapes, necessitating substantial classical computing resources. The demands increase nonlinearly as simulations acquire more detail. Quantum computing has long been proposed as a feasible approach to exceed classical simulation limits, but the transition from theory to practice has been hindered by the extensive number of qubits and circuit depth needed to accurately run even moderately complex scenarios without succumbing to errors.

      The OSSLBM algorithm directly addresses this challenge. It is built upon the quantum Lattice Boltzmann Method (QLBM), a well-regarded technique for translating classical fluid equations into quantum computing. The new framework minimizes the computational burden of each step, enabling a longer multi-step simulation to remain within the reliable execution capabilities of current quantum hardware.

      Haiqu’s middleware played a crucial role in this endeavor by reducing circuit depth, developing new algorithmic subroutines, and implementing specific error-reduction methods, allowing the system to carry out a workflow that would have been otherwise unattainable with today's devices.

      The importance of this achievement lies in the complexity of the obstacle. Earlier quantum CFD demonstrations primarily centered around simpler linear scenarios, focusing on fluid behavior without the interaction of a solid boundary. Understanding how fluid navigates around an object is essential for any industrially relevant application. Professor Oleksandr Kyriienko, Chair in Quantum Technologies at the University of Sheffield, characterized this work as “an interesting and timely contribution to quantum CFD,” emphasizing the necessity for further research in this area to develop industrially applicable quantum solutions.

      Quanscient and Haiqu have been working together on quantum CFD since at least 2024, when they reached the finals in the Airbus and BMW Quantum Mobility Challenge, and have previously showcased their efforts on IonQ hardware through Amazon Braket. Although industrial applications remain several years away, this current work represents a research milestone, affirming that their approach is viable on current hardware at this level of complexity.