Scaleway: CUDA-Q Integration Unifies GPU Supercomputing & European Quantum Hardware in Cloud
Summarize this article with:
Scaleway has unified GPU supercomputing with European quantum hardware, establishing a single cloud environment for developers working with both technologies. The company announced full compatibility of its Quantum as a Service platform with the NVIDIA CUDA-Q runtime, allowing users to write CUDA-Q kernels and execute workloads on either high-performance GPU clusters or actual quantum processors without altering their tools or managing separate systems. This integration provides direct access to Scaleway from within CUDA-Q, enabling state vector emulation on dedicated hardware scaling up to 38 qubits, surpassing the limits of many open-source emulators, and access to multiple quantum modalities. “CUDA-Q becomes a direct gateway to supercomputing scale infrastructure delivered as a cloud service,” Scaleway states, positioning itself as Europe’s leading multi-modal quantum cloud aggregator and bolstering the region’s quantum ecosystem. Scaleway and NVIDIA CUDA-Q Platform Compatibility Scaleway’s recent integration with the NVIDIA CUDA-Q platform establishes an environment where high-performance computing and quantum hardware converge within a single, sovereign cloud infrastructure. This compatibility allows developers to utilize familiar CUDA-Q kernels, directing workloads seamlessly to either Scaleway’s GPU clusters or quantum processors without requiring tool changes or complex environment management. This streamlined development process enables teams to write quantum programs once and then select their desired execution environment; large-scale simulations now run on NVIDIA Hopper and Blackwell architectures operated by Scaleway, or the same code can be deployed to real quantum processors via Scaleway’s network of partners. Valentin Macheret, Engineering Manager at Scaleway, said that “What we find particularly exciting about CUDA-Q is its multi-paradigm approach.” He explained that it brings classical HPC techniques and quantum development into the same workflow, which he believes is the right direction for the ecosystem. Scaleway’s platform currently supports simulations scaling up to 38 qubits, overcoming the memory limitations of many conventional systems, and offers access to quantum processors from IQM Quantum Computers and Alpine Quantum Technologies. Beyond emulation, the unified interface bridges experimentation and real-world validation, allowing rapid iteration from simulation to hardware. Sam Stanwyck, Director of Quantum Product at NVIDIA, noted that “Access to large scale simulations and quantum hardware are allowing researchers to explore and develop for future hybrid quantum-classical supercomputers.” By combining HPC-grade emulation with multi-modal quantum hardware, Scaleway positions itself as a leading European quantum cloud aggregator, strengthening the region’s quantum ecosystem and offering a secure, independent alternative for organizations prioritizing digital sovereignty. 38 Qubit State Vector Emulation via NVIDIA Blackwell This achievement bypasses limitations previously encountered around the 30-qubit threshold on conventional systems, effectively removing a major obstacle to larger-scale quantum simulations. The capability stems from utilizing eight interconnected NVIDIA Blackwell Ultra GPUs, linked via NVIDIA NVLink, to provide the necessary processing power and memory capacity for full state vector representation. Researchers previously constrained by hardware limitations can now instantly access supercomputing scale emulation without substantial capital investment or specialized infrastructure procurement. This on-demand platform is notable because it eliminates the need for users to provision GPU clusters, manage schedulers, or negotiate separate access to quantum hardware providers; CUDA-Q functions as a direct gateway to Scaleway’s cloud-delivered supercomputing infrastructure. Scaling simulations to 38 qubits is not merely an incremental improvement, but a substantial expansion of the accessible quantum space for experimentation and algorithm development. What we find particularly exciting about CUDA-Q is its multi-paradigm approach. It brings classical HPC techniques and quantum development into the same workflow, which feels like the right direction for the ecosystem. We’re looking forward not only to building on top of CUDA-Q, but also to contributing to it. Valentin Macheret, Engineering Manager at Scaleway This flexibility is achieved through partnerships with companies like IQM Quantum Computers, offering access to systems including Garnet (20 qubits), Sirius (16 qubits), and Emerald (54 qubits), alongside the 12-qubit IBEX Q1 trapped ion processor from Alpine Quantum Technologies. This expanded capacity isn’t merely an incremental improvement; it unlocks experimentation at a scale previously inaccessible without significant capital investment, as Valentin Macheret, Engineering Manager at Scaleway, highlights the strategic importance of this unified approach. Access to large scale simulations and state of the art quantum hardware are allowing researchers to explore and develop for future hybrid quantum-classical supercomputers Sam Stanwyck, Director of Quantum Product at NVIDIA Source: http://www.scaleway.com/ Tags: