Qilimanjaro Integrates QiliSDK with Nvidia CUDA-Q for GPU-Accelerated Quantum Emulation

[Andy89]

Spanish quantum computing company Qilimanjaro announced this week the integration of its QiliSDK software development kit with Nvidia's CUDA-Q platform, enabling GPU-accelerated quantum circuit emulation at scales previously inaccessible to its user base. The integration allows researchers and developers to run quantum circuit simulations on Nvidia GPU clusters through the CUDA-Q interface while using Qilimanjaro's native SDK for circuit design and optimisation. The practical result is that workflows developed in QiliSDK can be tested at larger qubit counts using classical GPU emulation before being deployed on actual Qilimanjaro quantum hardware.

Qilimanjaro specialises in analog quantum computing using superconducting flux qubits optimised for combinatorial optimisation problems. Its hardware approach differs from gate-based quantum computing: rather than executing discrete gate operations, analog quantum computers evolve a Hamiltonian that encodes an optimisation problem into the quantum system's ground state. This continuous approach is particularly suited to certain classes of problems including logistics scheduling, portfolio optimisation and protein folding, but makes it harder to emulate accurately on standard digital simulators. The CUDA-Q integration addresses this by providing GPU-accelerated simulation capabilities that can handle the continuous dynamics more efficiently.

Qilimanjaro also announced integration of its SDK with qBraid's cloud platform the same week, consistent with the pattern visible across quantum hardware companies of using qBraid as a middleware aggregation layer. The double announcement of CUDA-Q integration and qBraid access in the same week reflects a commercial maturation phase: quantum hardware companies are no longer just building hardware and publishing research, they are building the software ecosystem and cloud access pathways that enterprise customers need to develop and deploy applications.

Bull and Alice & Bob Sign New Agreement to Accelerate Europewww.hpcwire.com

[Joanne94]

Analog quantum computing getting GPU-accelerated emulation through CUDA-Q solves a specific and real problem: the simulation techniques for continuous Hamiltonian evolution are different from gate-based circuit simulation and classical tools for them are much less developed

[RicFlair_X]

Qilimanjaro being Spanish and competing with French, Finnish, and American quantum companies in the same European cloud infrastructure space illustrates how the European quantum ecosystem has developed genuine breadth across multiple countries with distinct technical approaches

[Cheeky Blake]

The combinatorial optimisation focus for analog quantum computing is where the near-term commercial case is strongest. Logistics, portfolio optimisation and scheduling are all problems where quantum devices are already being evaluated against classical solvers in real enterprise contexts

[Maya98]

CUDA-Q integration appearing in announcements from Qilimanjaro, Alice and Bob, Quandela, Pasqal and qBraid in the same week confirms that Nvidia has executed its quantum ecosystem strategy successfully. The entire European quantum stack is being wired to run through CUDA-Q

[Kieron83]

qBraid as the middleware layer aggregating multiple quantum hardware providers is the neutral broker role that has value precisely because it is not hardware-specific. If you build on qBraid's API you can switch between Qilimanjaro, IBM, IonQ and others without rewriting your code

[ECWAlfie47]

The analog versus digital quantum computing distinction matters for near-term applications. Analog devices are optimised for specific problem classes and can outperform digital quantum computers on those classes today. Digital devices are more general purpose but currently noisier for most tasks

[HitmanBrad98]

Enterprises evaluating quantum for optimisation problems want to test quantum versus classical before committing to hardware. GPU-accelerated emulation at scale is what enables that benchmarking. Without emulation access the evaluation process requires expensive hardware time that most organisations are not ready to purchase