NVIDIA’s CUDA-Q platform joins SUSE’s AI Factory

Introduction

SUSE’s AI Factory and NVIDIA’s CUDA‑Q platform are being combined with Fujitsu’s server hardware to create a unified, enterprise‑grade hybrid quantum computing stack. Announced in October 2026, the solution lets IT teams treat quantum circuits as ordinary workloads, extending AI‑driven orchestration into the quantum domain while keeping data and algorithms on‑premises.

A seamless bridge between classical AI and quantum workloads

The integration embeds CUDA‑Q directly into Fujitsu’s servers, allowing the AI Factory to schedule, monitor, and manage quantum tasks using familiar tools and workflows. Key benefits include:

• Unified lifecycle management – CUDA‑Q is packaged and version‑controlled by the AI Factory, eliminating manual patching.
• Standardized deployment – Reference architectures and a single support contract cover both the hardware and software layers.
• On‑premises sovereignty – All quantum simulations and data remain within the customer’s private infrastructure, preserving data residency and security.

How the stack works in practice

1. Define a quantum circuit using CUDA‑Q’s Python‑like API.
2. Submit the circuit through the AI Factory’s compute broker, which routes the job to the appropriate quantum simulator or hardware accelerator on the Fujitsu platform.
3. Execute and retrieve results – The runtime handles error‑mitigation, calibration data, and any required classical post‑processing, returning results to the originating application.
4. Integrate with AI services – Built‑in modules such as Retrieval‑Augmented Generation (RAG) and LLM serving let developers combine quantum‑enhanced inference with existing AI models.

What this means for enterprises

• Leverage existing investments – Companies can augment current AI pipelines with quantum acceleration without overhauling their data‑center architecture.
• Accelerate R&D – Researchers gain production‑ready access to quantum simulation, shortening the path from prototype to deployment.
• Simplify operations – A single stack reduces the complexity of managing disparate quantum and classical components, lowering total cost of ownership.

Looking ahead

The hybrid stack is slated for a late‑2026 release, positioning it as one of the first production‑ready quantum‑classical solutions for sovereign enterprises. As Fujitsu expands its portfolio with new quantum processors and the Digital Annealer, the partnership promises to broaden the range of workloads that can benefit from quantum speed‑ups while maintaining strict security and compliance standards.

--- The integration of NVIDIA’s CUDA‑Q with SUSE’s AI Factory marks a significant step toward making quantum computing a practical, manageable part of enterprise AI infrastructure.