Qiskit beyond Python: Fortran, C++, and Julia

IBM Research is pushing Qiskit past its Python comfort zone. Since Qiskit v2.0, the quantum SDK has exposed a C API to its Rust core, and that one interface now opens native paths from Fortran, C++, and Julia without forcing Python into the middle.

That is more than a packaging trick. IBM says all of the bindings share the same Qiskit shared library, which means they can interoperate. In theory, a circuit can be built in Fortran and handed off to C++. For teams in high-performance computing, that matters because the rest of the application is often already written in those languages, while Python is just glue.

The Fortran route, qiskit-fortran, uses the standard iso_c_binding interface to call the C API directly. IBM says that keeps the workflow native, avoids Python objects and the GIL, and lets existing Fortran memory flow straight into circuit construction. The C++ path, qiskit-cpp, is header-only and links against Qiskit’s C library. Julia gets perhaps the smoothest entry: Qiskit.jl wraps the API in native Julia types, runs in Jupyter notebooks, and comes with QiskitIBMRuntime.jl for sending jobs to IBM Quantum hardware.

IBM’s bigger pitch is quantum-centric supercomputing: quantum calls as linked subroutines inside classical codes, not a separate sidecar process. That fits the company’s examples. The article points to workflows in optimization, machine learning, and simulation, then walks through a Julia notebook that models the Schrödinger equation on a transverse-field Ising chain. At 100 qubits, the point is blunt: classical reference methods start to strain, while the hardware keeps the entanglement alive.

The practical message is clear. IBM wants quantum computing to meet scientific computing where it already lives, in Fortran, C++, and Julia, not drag everyone through a Python detour. That is the right instinct. Researchers rarely ask for a fresh interpreter; they ask for less friction, and the old HPC languages are where the actual friction lives.