Qilimanjaro's QiliSDK 0.2.0 Pushes GPU-Accelerated Quantum Emulation Past 30 Qubits With Unified Digital-Analog Framework

[Vector14]

Spanish quantum computing company Qilimanjaro released QiliSDK 0.2.0 on June 23, an open-source software development kit that unifies digital and analog quantum programming through a single backend-agnostic Python framework. The release integrates Nvidia's CUDA-Q platform for GPU-accelerated emulation, pushing state-vector simulation capability past 30 qubits, a meaningful threshold given that classical simulation of quantum systems grows exponentially more demanding with each additional qubit. The SDK has already been deployed across multiple European supercomputing centres.

The unification of digital and analog programming paradigms within one framework addresses a genuine fragmentation problem in the quantum software ecosystem. Digital quantum computing executes discrete gate operations on qubits, the model most commercial quantum hardware and most existing software tooling is built around. Analog quantum computing, which Qilimanjaro's own hardware uses, evolves a continuous Hamiltonian that encodes an optimisation problem directly into the quantum system's physical dynamics, an approach particularly well suited to combinatorial optimisation problems including logistics scheduling and portfolio construction but historically poorly served by simulation tools designed around discrete circuit models. QiliSDK's three-tier architecture and comprehensive noise modelling are designed to let researchers prototype and test both paradigms within the same codebase rather than maintaining separate toolchains.

Interoperability was a stated design priority, with the release supporting both OpenQASM 3, the industry-standard intermediate representation for quantum circuits, and Microsoft's Quantum Intermediate Representation. This positions QiliSDK as infrastructure that can sit underneath workflows targeting multiple hardware backends rather than locking developers into Qilimanjaro's own analog quantum hardware exclusively. The release follows Qilimanjaro's integration announcements with both Nvidia CUDA-Q and the qBraid cloud aggregation platform in the same week, a pattern of simultaneous multi-platform integration that has become characteristic of the European quantum software ecosystem's commercial maturation through 2026.

Qilimanjaro Releases Qilisdk 0 2 0 And Integrates Nvidia Cuda Q For Unified Hybrid Hardware Emulationthequantuminsider.com

[alwaysPatrick19]

Pushing GPU-accelerated state-vector simulation past 30 qubits matters because that is genuinely close to the boundary where classical simulation starts becoming computationally prohibitive even with significant GPU acceleration. Every additional simulated qubit roughly doubles the memory and compute requirement