Forschungszentrum Jülich request: Qiskit 2.x and SciPy>=1.16.x support.

[AlvaroHuanay]

Dear Qiskit Machine Learning team,

First of all, I want to thank you for the excellent work on this library. We are contacting you from Forschungszentrum Jülich. I noticed that in the current release (0.8.4) the requirements.txt still pins: qiskit>=1.0,<2.0 and scipy>=1.4,<1.16

With Qiskit 2.x now released and SciPy 1.16.x available, this pin prevents installation in environments using the latest versions. I would like to request for a loosening in the requirements, specifically: Qiskit ≥2.0 and SciPy ≥1.16 (since we support SciPy 1.16.1).

I am aware of PR #973 , which appears to be migrating the library to Qiskit 2.0. If this issue is already tracked there, please consider this a +1 request for SciPy ≥1.16 support as well.

Wishing you a nice day.

Thanks a lot!

[edoaltamura]

Hi Alvaro! Well spotted, we are indeed removing the constraint on the Scipy version in #973, when we migrate to Qiskit 2.0+. Scipy 1.16 seemed to break some nightly CI tests a few weeks ago, hence the pinning. Did you encounter any issues with 1.16 while running locally?

[AlvaroHuanay]

Hi @edoaltamura,

Thanks for the quick reply. We haven’t run into any issues with SciPy 1.16 on individual Qiskit (v1.4.4) installations. However, we’re preparing a cloud service that integrates several quantum SDKs (Pulser, Qutip, DWave, Qiskit...) so we need to maintain a common set of dependencies across all of them (plus other non-quantum SDKs).

At the moment, we can’t move away from SciPy 1.16.1 because it’s already in production for Stages/2026, which is why I opened this issue.

Since the SciPy pin will be removed after the Qiskit 2.0+ migration, we’re planning to wait for that release. Do you have an estimate for when the migration will be completed? This would help us coordinate with other developers, as our quantum cloud service is planned for release by December 2025.

Thank you very much again for your help!