SUSE and Fujitsu partner to bring quantum workloads into enterprise data centers

SUSE and Fujitsu are promising a production-ready AI and quantum platform that ordinary enterprise IT teams can manage. Their October 21, 2026 partnership announcement sets availability for late 2026, however. The commitment is to deliver an integrated system for enterprise data centers, not a platform customers can already deploy.

The planned system brings together SUSE AI Factory, NVIDIA AI Enterprise and NVIDIA CUDA-Q on Fujitsu servers. Its central idea is to treat quantum circuits as extensions of workloads the enterprise platform already handles, rather than manage them through a separate operational setup.

The companies say on-premises or private-infrastructure deployment will keep proprietary intellectual property, algorithms and enterprise datasets within customers’ infrastructure. That is the practical meaning of the announcement’s sovereignty pitch: preserving data residency and local security control.

Running this stack on Fujitsu servers allows customers to supercharge their existing hardware investments and connect quantum simulation directly to production applications.

Udo Wuertz, CTO Fsas Technologies EMEA at FSAS

Software upkeep, circuit execution and one support contract

SUSE says administrators, cloud architects and platform engineers will be able to use familiar enterprise tools and security policies without specialized quantum physics expertise. The announced architecture assigns three jobs to the software:

• Software upkeep: SUSE AI Factory will package, deploy and manage CUDA-Q throughout its lifecycle. SUSE says this will remove manual patching and the work of matching compatible software versions.
• Reusable AI components: A modular building-block layer includes retrieval-augmented generation, agentic coding and language-model serving. The announcement describes this layer as standardizing AI deployment across Fujitsu and Supermicro hardware.
• Workload execution: CUDA-Q will connect directly to the compute broker, which coordinates computing work. The companies say that integration will automate circuit submission and execution for hybrid quantum workloads.

Support is also bundled. The proposal includes pre-validated reference architectures—designs checked in advance—and a single support contract covering the hardware and software stack. The companies say that arrangement, together with automated software management, will reduce setup time and operational complexity.

These are operational promises, not measured computing gains. The announcement provides no benchmark results demonstrating faster workloads or better business outcomes.