Preparing for Hybrid HPC and Quantum Computing

What happens when emerging quantum computing resources need to work alongside the high-performance computing (HPC) infrastructure organizations already depend on?

HPC remains the foundation for large-scale simulation, AI, data processing, and other compute-intensive workloads. Quantum computing is emerging as a complementary resource for selected computational tasks, working alongside classical systems rather than replacing them. This creates a new infrastructure challenge. Organizations need ways to connect HPC, quantum processors, and other computing resources within coordinated workflows.

The goal is not to predict when quantum computing will become mainstream. It is to prepare existing HPC environments to accommodate new computing resources as applications mature. 

This article examines how HPC infrastructure, orchestration, and workflow management can support hybrid quantum-classical computing.

How Quantum Computing is Shaping HPC Planning

As quantum computing moves closer to practical applications, HPC teams need to consider how these emerging resources could fit into existing computing environments. The challenge is not replacing HPC, but preparing infrastructure that can accommodate quantum resources alongside established CPU and GPU systems.

Quantum hardware and software are advancing, but broader organizational readiness is still developing. The lack of quantum readiness is also visible outside HPC. ISACA found that 55% of organizations had not taken steps to prepare for quantum computing in cybersecurity, while only 5% reported having a defined quantum computing strategy. 

For HPC teams, this reinforces the value of planning for quantum integration before specific workloads require it.

Assessing Workloads for Future Quantum Integration

Organizations can evaluate workloads that may eventually benefit from quantum acceleration without immediately purchasing quantum hardware. Hence, early planning should focus on:

• Identifying potential quantum use cases

• Understanding existing HPC workloads and dependencies

• Assessing data preparation and transfer requirements

• Reviewing current scheduling and orchestration capabilities

• Building skills around quantum-classical workflows

This approach creates a path toward hybrid computing without requiring a complete infrastructure redesign.

HPC Will Remain Part of the Quantum Workflow

Quantum processors are expected to operate alongside classical computing resources. Classical systems can handle tasks such as data preparation, problem formulation, preprocessing, and post-processing. Quantum processors, by contrast, handle specific computational tasks within a broader quantum-classical workflow. 

Recent projects are already demonstrating this approach in practice. For example, D-Wave and Nasdaq Verafin are exploring quantum-hybrid computing to support predictive systems that detect unusual account behavior tied to fraud and money laundering. This shows how quantum and classical resources can integrate within a broader computing environment.

This means HPC infrastructure remains relevant even as quantum capabilities develop.

Building Hybrid Readiness Takes Time

Organizations can build readiness gradually rather than waiting for mature quantum applications. Instead of treating quantum readiness as a separate infrastructure project, teams can strengthen capabilities that already support complex HPC environments and can later accommodate new resource types. Key capabilities include:

• Flexible workflow design

• Reliable data pipelines

• Resource-aware scheduling

• Secure connectivity

• Monitoring across heterogeneous resources

• Skills for managing hybrid workloads

These capabilities provide the foundation for the next stage: coordinating classical and quantum resources within the same workflow.

What Changes When HPC and Quantum Work Side by Side

A hybrid environment changes how organizations think about workload execution. HPC will continue to handle most general-purpose computing, while quantum resources can be assigned to specific tasks within larger workflows.

A single application may therefore move through several stages of computation. Data can be prepared on classical infrastructure, processed by a quantum resource, and then returned to classical systems for analysis. 

Research shows this pattern. One example is RIKEN's closed-loop Fugaku workflow, where classical and quantum execution stages feed into one another. Commercial platforms follow the same pattern. Altair, now part of Siemens, built a hybrid classical-quantum platform with partner G2Q for applications such as fraud detection.

This pattern of separate classical and quantum execution stages introduces additional planning requirements. 

Current State vs. Hybrid-Ready State

Planning requirements become clearer when you compare the current environment and a hybrid-ready environment side by side. The table below summarizes how these requirements differ between traditional HPC and a hybrid-ready environment.

Connectivity also becomes more important. Quantum processors may be located outside the primary HPC environment or accessed through remote services. This introduces considerations around network latency, availability, access control, and data transfer.

Hybrid systems therefore require more than another compute resource. They require infrastructure that can coordinate different resource types and move workloads between them.

How Orchestration Connects HPC and Quantum Workflows

Orchestration provides the control layer needed to coordinate heterogeneous computing resources. Instead of treating quantum computing as an isolated environment, orchestration can connect quantum tasks to the surrounding classical stages. 

For HPC teams, this makes orchestration important in three practical areas: deciding where workloads should run, connecting interdependent stages, and maintaining visibility across the full workflow.

Route Work to the Right Resource

Applications are designed and partitioned so certain computational stages run on classical infrastructure while others suit a quantum resource. This partitioning happens at the application and algorithm level, based on which parts of the problem quantum hardware can address.

Once that partitioning is defined, the orchestration layer coordinates how those stages actually execute. This can help manage:

• CPU and GPU workloads

• Quantum processing tasks

• Dependent workflow stages

• Resource availability

• Execution priorities

The scheduler or orchestration layer coordinates these stages and resources rather than simply launching individual jobs. It manages execution, not the algorithmic decision of which computational logic belongs on the quantum processing unit (QPU).

Connect Data Across Workflow Stages

Hybrid workflows may involve several connected stages, including classical data preparation, quantum execution, and classical post-processing. Reliable data movement is essential when these stages depend on one another. 

The workflow must transfer the required inputs, collect quantum results, and make those results available to subsequent classical tasks. Research into hybrid quantum-HPC systems also highlights the importance of communication and latency between quantum and classical processors.

Maintain Visibility and Control

Hybrid environments add more resources and dependencies to monitor. Organizations need visibility into:

• Workload status

• Resource availability

• Workflow dependencies

• Data movement

• Quantum job execution

• Performance and utilization

• Access and governance

This visibility helps teams identify bottlenecks and understand where time is spent across the entire workflow.

How Organizations Can Prepare for Hybrid HPC and Quantum Computing 

Preparing for hybrid HPC and quantum computing starts with practical steps. Organizations can build on their existing HPC capabilities without investing in quantum hardware immediately. The steps below outline how to prepare for this shift:

1. Assess Potential Workloads

Review existing applications for tasks that quantum hardware is best suited for, such as optimization problems, pattern detection, and analysis of large or complex datasets like fraud detection. For each workload, identify which stages need to stay on classical systems. Preprocessing, data preparation, and result validation will continue to run on existing HPC resources regardless of where the quantum stage fits in.

2. Build Flexible Workflows

Design workflows so the quantum stage can be swapped out rather than fixed in place. This means using the same workflow tools for quantum jobs as for classical jobs, so you can add or change a quantum resource later without rebuilding the surrounding pipeline. This is important because quantum hardware, software, and access models are still changing quickly.

3. Plan Hybrid Infrastructure

Extend existing scheduling and orchestration systems to treat quantum resources as another resource type, with their own access controls and availability, rather than building a separate system. This includes adding network connections and access controls for remote or cloud-based quantum resources, and implementing monitoring that can follow a job as it moves between classical and quantum systems.

4. Build Readiness Gradually

Start with a research project or pilot that runs on real infrastructure, so integration issues like data transfer and job coordination surface early. Use that experience to test the workflow and infrastructure choices above before moving to larger deployments.

The objective is not to build a separate quantum environment. It is to make your existing HPC environment flexible enough to incorporate new computing resources when they provide practical value.

Early Hybrid Applications Show What Future Infrastructure May Require

Early applications provide a useful view of how quantum computing may operate alongside HPC. Fraud detection is one example. ADOPT-TEM is a National Quantum Computing Center project led by Algorithmiq UK with partner Rigetti UK. It worked to improve the hardware performance of a hybrid quantum-classical fraud detection method. The project also tested how the method extends into a quantum + HPC workflow on NQCC infrastructure.

This model is also appearing in commercial HPC software. Siemens has demonstrated hybrid quantum-classical workflows for fraud detection. These fraud-detection workflows use HPCWorks™ software components for scheduling and workflow orchestration. 

HPCWorks includes explicit support for hybrid quantum-classical workflows, and Siemens identifies complex pattern detection and fraud detection as early use cases.

These examples highlight several infrastructure requirements:

• Integration between classical and quantum resources

• Coordinated workload execution

• Reliable data movement

• Classical preprocessing and post-processing

• Workflow orchestration

• Resource management

• Connectivity and monitoring

The lesson is that hybrid applications depend on the complete workflow, not simply on access to a quantum processor.

How HPCWorks Supports Hybrid Compute

The infrastructure needed for hybrid quantum-classical computing builds on capabilities many HPC teams already use. HPCWorks provides HPC and cloud orchestration that manages workloads and resources across computing environments and provides a foundation to support emerging hybrid compute use cases.

Workload Orchestration

HPCWorks adds AI-assisted scheduling features, including memory resource prediction, to streamline job submission and optimize resource use across CPU and GPU workloads. This scheduling foundation is what Siemens extends to run hybrid quantum-classical workflows for pattern-detection use cases such as fraud detection.

Heterogeneous Resource Management

Future HPC environments may combine CPUs, GPUs, quantum processors, cloud resources, and other accelerators. Managing these resources through coordinated infrastructure can reduce the need to treat every resource as a separate environment.

Interoperability and Connectivity

Quantum resources may be deployed locally, at research facilities, or through cloud services. Hybrid infrastructure therefore needs mechanisms for connecting workloads to resources outside the immediate HPC cluster.

This approach supports gradual evolution. Organizations can continue using their existing HPC infrastructure while adding new resource types as applications become technically and economically practical.

Conclusion

Quantum computing is unlikely to replace HPC. The more practical direction is coexistence, where quantum processors handle selected tasks while classical HPC continues to manage data preparation, simulation, processing, and post-processing. Preparing for this model requires more than access to a quantum processor. Organizations need flexible workflows, reliable data pipelines, heterogeneous resource management, connectivity, and orchestration.

Organizations can begin preparing today through workload assessment, workflow modernization, integration planning, and targeted experimentation. As quantum applications mature, existing HPC environments can evolve with them. The organizations best positioned for hybrid computing will treat quantum as another resource within a coordinated computing architecture, not as a separate technology stack.

For engineering and research teams preparing for heterogeneous compute environments, contact Siemens to see how HPCWorks can support flexible workflows and emerging quantum-classical applications.