ORCA Computing, in collaboration with the Poznań Supercomputing and Networking Center (PCSS), has unveiled a groundbreaking blueprint for integrating photonic quantum AI into operational data centers. By combining ORCA’s photonic quantum processors with NVIDIA’s accelerated computing and CUDA-Q software platform, this initiative demonstrates a scalable, deployment-ready approach for hybrid quantum–classical systems, advancing next-generation AI and machine learning workloads.
ORCA Computing and PCSS create quantum AI data center blueprint.
Integrates photonic quantum processors with NVIDIA Hopper GPUs.
Uses NVIDIA CUDA-Q for seamless quantum–classical programming.
Demonstrated in PCSS data center with Imperial College London.
Enables scalable hybrid machine learning and HPC applications.
Marks first photonic quantum integration via CUDA-Q platform.
ORCA Computing, alongside PCSS and Imperial College London, has successfully demonstrated a distributed photonic quantum neural network within a real-world data center. The system integrates two ORCA PT photonic quantum processors with NVIDIA Hopper GPUs, housed in standard 19-inch rack cabinets. Operating at room temperature and managed through the NVIDIA CUDA-Q platform and Slurm workload manager, this setup allows seamless programming of classical and quantum resources. This marks the first integration of photonic quantum processors via CUDA-Q, establishing a scalable framework for hybrid machine learning and high-performance computing (HPC) applications.
The collaboration leverages award-winning research from Imperial College London, recognized as the Best Paper in the photonics track at a premier quantum conference during IEEE Quantum Week. Conducted in PCSS’s operational data center, the demonstration validates the viability of quantum-accelerated architectures without requiring extensive infrastructure changes. As Krzysztof Kurowski, Deputy Director & CTO of PCSS, stated: “The open, easy integration of multiple QPUs and NVIDIA GPUs demonstrates that quantum accelerated supercomputing is not just tomorrow's promise but today's reality.” This effort builds on ORCA’s work with the UK’s National Quantum Computing Centre, further solidifying its leadership in hybrid quantum initiatives.
The blueprint showcases a practical path for organizations to adopt quantum-enhanced systems. By combining ORCA’s photonic quantum processors with NVIDIA’s GPU technology, the system supports advanced AI and machine learning workloads while maintaining compatibility with existing data center infrastructure. Sam Stanwyck, Group Product Manager for Quantum at NVIDIA, noted: “The integration of quantum into the fabric of HPC is a critical step towards developing the future of quantum accelerated supercomputing, combining multiple quantum accelerators with many classical accelerators.” The CUDA-Q platform’s open and performant design accelerates progress toward scalable quantum solutions.
This milestone moves quantum computing from theoretical concepts to practical HPC environments. ORCA and its partners plan to enhance algorithmic capabilities, reduce latency, and scale to next-generation quantum systems, paving the way for large-scale, impactful quantum computing. The initiative aligns with ORCA’s global deployments, including systems at Montana State University and the UK Ministry of Defence, reinforcing its role in advancing quantum AI and HPC integration worldwide.
ORCA Computing, headquartered in London, UK, with offices in the United States, is a leading developer and provider of full-stack photonic quantum computing systems. The company delivers an innovative approach to quantum computing, providing robust, high-performance, and data center-standard systems for machine learning, generative AI and optimization workloads. ORCA Computing has successfully delivered ten on-premises quantum computers to leading global customers, including the UK National Quantum Computing Centre, Montana State University, and the Poznan Supercomputing and Networking Center.