Artificial Intelligence, Quantum Hardware

Quandela and NVIDIA link quantum processors to AI with NVQLink

Avatar of Ivy Delaney
September 14, 2026 by Ivy Delaney

At IEEE Quantum Week 2026 in Toronto, Quandela is demonstrating a pathway to integrate quantum processing units (QPUs) into existing artificial intelligence and high-performance computing infrastructure using NVIDIA NVQLink, the company says. The companies are publishing a technical white paper outlining an architecture where GPUs remain central to AI workflows, with the QPU acting as a specialized accelerator.

“The challenge is no longer just about accessing a QPU, but about integrating it directly into AI workflows,” says Jean Senellart, Chief Technology and Product Officer at Quandela, as the collaboration aims to identify and scale quantum applications without requiring a complete overhaul of current computing environments.

Photonic QPU Integration with NVIDIA NVQLink for AI Workflows

NVIDIA’s January 2026 launch of NVQLink established a sub-microsecond latency connection between NVIDIA GPUs and quantum system controllers, a critical element in enabling practical hybrid quantum-classical workflows. This low-latency communication is now being demonstrated by Quandela, integrating its photonic quantum processing unit (QPU) with NVIDIA’s infrastructure to accelerate specific AI workloads, according to the company. This integration isn’t about replacing existing computing infrastructure, but augmenting it, according to the published architecture.

CPUs orchestrate overall workflows, and GPUs continue to handle intensive computation and simulation tasks vital for AI development. The photonic QPU then steps in to address specific computational challenges where quantum algorithms offer a potential advantage, allowing researchers to explore and validate hybrid algorithms, particularly in quantum machine learning.

NVIDIA’s CUDA-Q platform and cuQuantum SDK facilitate algorithm development, simulation, and testing on GPUs before deployment on the photonic QPU, creating a streamlined development pipeline. This integration allows users to progress incrementally toward quantum computing, using existing infrastructure to identify high-potential applications before scaling to more powerful quantum systems.

Sam Stanwyck, Director of Quantum Product at NVIDIA, explains that “Quantum computing becomes more useful when it moves beyond standalone access and becomes part of the accelerated computing researchers already use.” This collaborative effort aims to bridge the gap between quantum potential and practical AI applications, offering a pathway for researchers to explore and deploy quantum-enhanced solutions within their current workflows.

The challenge is no longer just about accessing a QPU, but about integrating it directly into AI workflows. With NVIDIA, we are opening a path where the photonic QPU becomes a specialized accelerator for exploring and validating hybrid algorithms, particularly in QML, before scaling the most promising use cases.

Sam Stanwyck, Director of Quantum Product at NVIDIA

CUDA-Q and cuQuantum Enable Algorithm Development & QPU Simulation

NVIDIA’s CUDA-Q platform and cuQuantum SDK are now central to a collaborative effort with Quandela, enabling algorithm development and simulation prior to deployment on photonic quantum processing units. This pre-deployment testing, facilitated by GPU acceleration, allows researchers to refine quantum algorithms and identify optimal workloads for quantum advantage. The companies report that this approach bypasses the need for entirely new infrastructure, using existing HPC environments to explore quantum applications.

The integration detailed in a technical white paper released builds on NVIDIA NVQLink, a low-latency connection between the GPU and the Quantum System Controller that manages the QPU. NVIDIA’s 2023 partnership with Rigetti Computing, naming the company a CUDA-Q hardware partner, demonstrates a broader strategy of supporting diverse QPU architectures within its ecosystem, the firm reports. Further collaborations with QuEra Computing and IQM, established in 2025, have integrated their processors with NVQLink, expanding the range of accessible quantum hardware.

This step-by-step model, presented at IEEE Quantum Week 2026, emphasizes access, integration, and scale as key phases in quantum computing adoption. According to the companies, the initial “Access” phase allows experimentation without requiring dedicated quantum infrastructure. The focus is on embedding the QPU within existing CPU and GPU environments to develop and test algorithms, pinpointing areas where quantum computing can deliver tangible benefits. The integration with Classiq’s quantum software platform, completed in March 2026, further streamlines algorithm development and deployment within the CUDA-Q framework.

Three-Step Approach: Access, Integration & Discovery, and Scale

The three-step model guiding Quandela and NVIDIA’s integration of quantum processing units centers on a phased approach, Access, Integration & Discovery, and Scale, designed to minimize disruption to existing computational infrastructure. This progression prioritizes practical application over wholesale replacement, allowing organizations to explore quantum capabilities without immediate, large-scale investment. The initial “Access” phase facilitates experimentation by providing remote access to Quandela’s photonic quantum processing unit, bypassing the need for dedicated on-site quantum hardware.

The culmination of this process, “Scale,” aims to expand applications demonstrating potential, moving toward larger and more powerful quantum systems. “Our work with Quandela provides a step-by-step model for bringing photonic QPUs into NVIDIA quantum-GPU supercomputing, helping teams identify promising applications today and scale them as more powerful systems come online,” the company reports.

This strategy aligns with NVIDIA’s broader investment in quantum computing, including partnerships with Quantinuum, QuEra Computing, and IQM, all integrated with its CUDA-Q platform. Quandela’s teams will discuss the integration of photonic QPUs into AI and HPC infrastructures with attendees, showcasing a vision for a hybrid computing future. This presentation signals a commitment to transparency and collaborative development within the quantum computing community, offering a concrete demonstration of the technology’s potential.

Quantum computing becomes more useful when it moves beyond standalone access and becomes part of the accelerated computing researchers already use. Our work with Quandela provides a step-by-step model for bringing photonic QPUs into NVIDIA quantum-GPU supercomputing, helping teams identify promising applications today and scale them as more powerful systems come online.

Sam Stanwyck, Director of Quantum Product at NVIDIA
More like this
Quantum Research NewsIonQ and partners show practical applications and future roadmap at IEEE 2026
Artificial IntelligenceMolybdenum disulfide transistors reshape spiking neural networks
Quantum HardwareLogical advantage of Pinnacle confirmed on spin-qubit hardware
Artificial IntelligenceAnthropic Proposes 4 Ways to Ensure Democratic Oversight of AI Progress
Stay current

See today’s quantum computing news on Quantum Zeitgeist for the latest breakthroughs in qubits, hardware, algorithms, and industry deals.

Tags:
CUDA-Q GPU GPUs HPC NVIDIA QPU quandela Quantum Computing
Avatar of Ivy Delaney

Ivy Delaney

Ivy Delaney has been working with neural networks and machine learning since the mid-nineties, back when a couple of hidden layers and a long afternoon of training counted as ambitious. She has watched the field go from academic curiosity to the thing quietly running underneath everything, and she brings that long view to quantum computing. For Quantum Zeitgeist she covers the ground where the two fields meet. That means quantum machine learning and the variational algorithms it leans on, and it also means the less glamorous but more interesting story of classical machine learning already doing real work inside quantum machines, decoding error-correcting codes, calibrating noisy hardware and learning the error models that simulators depend on. She writes about the hardware those algorithms have to run on too, and about the post-quantum cryptography scramble that the same hardware has set off. Her stories typically start with the paper, whether that is peer-reviewed work, conference proceedings or an arXiv preprint, with the source linked so you can hold a claim up against the research it came from. She is unimpressed by benchmarks that will not say what they beat, and by demonstrations that only work in the press release.

Latest Posts by Ivy Delaney: