As the HPC community prepares for SC26 and its “HPC Unites” theme, quantum computing is increasingly part of the conversation about how supercomputing resources get used. Classiq provides a software engineering platform for building quantum applications and connecting them to the wider computing environment. Developers describe algorithms in its high-level Qmod language or Python SDK, and a synthesis engine generates optimized circuits for the chosen hardware and resource constraints. For SC attendees, Simon Fried, VP of Corporate Communications, likens it to design automation applied to quantum computing. The past year brought notable news: Intel announced in October its agreement to acquire Classiq, with the transaction expected to close in the first half of 2027, subject to regulatory approvals and other customary conditions.
In this VMblog pre-show Q&A, Fried discusses how organizations can build quantum development capabilities around the accelerated computing infrastructure they already have, and he offers concrete milestones for CFOs and CIOs under budget scrutiny. He walks through collaborations with NVIDIA, Oracle, Comcast, and AMD, including a demonstrated cut in a 31-qubit CUDA-Q workflow from 67 minutes to 2.5 minutes, while being careful to note that this is a simulation result and not evidence of quantum advantage. He also covers how AI agents are changing quantum development, the governance needed around production HPC environments, the role of the Fault Tolerance Engine, the deployment and data residency questions shaping procurement, and the bets Classiq is making looking toward SC27.
++
VMblog: If you were giving an SC26 attendee a quick overview of your company, what would you say? How has your company evolved since SC25, and what has been the biggest change in your business over the past twelve months?
Simon Fried: Classiq provides the software engineering platform for developing quantum applications and connecting them to the wider computing environment. Developers describe algorithms in Qmod, our high-level language, or through our Python SDK. Our synthesis engine generates optimized quantum circuits for the chosen hardware and resource constraints. For an SC attendee, the familiar idea is design automation applied to quantum computing.
Since SC25, we have expanded AI-assisted development, accelerated our NVIDIA CUDA-Q workflow and introduced our Fault Tolerance Engine. The biggest corporate development is Intel’s agreement to acquire Classiq, announced in October. The proposed combination brings our quantum software synthesis together with Intel’s silicon spin-qubit technology and manufacturing expertise. The transaction is expected to close in the first half of 2027, subject to regulatory approvals and other customary conditions. Until then, the companies remain independent. Intel has stated its intention to continue supporting Classiq’s multi-platform integrations and hardware-agnostic approach after closing.
VMblog: What key message will SC26 attendees hear from you this year? With AI infrastructure budgets under heavy scrutiny and hardware costs rising, what concrete value proposition can they take back to their CFO and CIO?
Fried: Our message is that organizations can build useful quantum development capabilities around the accelerated computing infrastructure they already have. The immediate value comes from shorter development cycles, more efficient circuits and better evidence for deciding where to invest.
One concrete example is our NVIDIA CUDA-Q integration. In a demonstrated 31-qubit workflow on a single NVIDIA A100 GPU, the time from circuit synthesis through completed simulation fell from 67 minutes to 2.5 minutes, roughly a 27-fold speedup. That gives researchers considerably more opportunity to test and refine an approach. It is a workflow-specific simulation result, rather than evidence of quantum advantage over a classical application solver.
For the CFO and CIO, we recommend measurable milestones. Track engineering time, compute consumption, solution quality against a credible classical baseline and the resources a future quantum implementation would require. Those measures help an organization build expertise while keeping investment tied to evidence.
VMblog: This year’s SC theme is “HPC Unites.” Where does your company fit into that idea? Which partnerships, joint reference architectures, or open standards are you announcing or demonstrating that make the whole greater than the sum of its parts?
Fried: Classiq connects application development with the processors and infrastructure needed to test and execute quantum programs. That is a natural fit with ‘HPC Unites,’ because a useful quantum workflow depends on close cooperation across the computing stack.
Our NVIDIA CUDA-Q integration connects high-level quantum models to accelerated hybrid execution. With Oracle, we demonstrated a 36-qubit portfolio optimization simulation across eight NVIDIA A100 GPUs on OCI. With Comcast and AMD, we developed a quantum approach to resilient network routing, using Classiq to generate the circuits and Qiskit Aer on AMD Instinct MI300X GPUs to simulate them. AMD is also an investor in Classiq.
These collaborations bring domain expertise, software engineering and accelerated computing into the same development process. They give HPC teams concrete workflows to examine and extend, with applications ranging from finance to telecommunications and engineering simulation.
VMblog: Agentic AI is now on the SC program, including a dedicated workshop. How are AI agents changing your product, whether in scheduling, code optimization and migration, workflow orchestration, or system operations? What governance and access controls do you put around agents that touch production HPC environments?
Fried: Our AI agents help developers translate a problem or algorithm into a structured quantum program, refine the implementation and prepare experiments. They work with Qmod and our synthesis engine, so the output can be compiled, inspected and tested within the development workflow.
In the Oracle proof of concept, an agent produced a portfolio optimization notebook in under 15 minutes. The team then reviewed it, adjusted parameters and validated execution before running the larger simulation. That is a useful example of how AI can reduce the effort needed to reach a testable application.
For production HPC environments, our recommended approach is to keep execution under the operator’s established controls. Agent-assisted development should use scoped credentials, explicit access to approved resources and review before costly or sensitive jobs run. Compilation checks alone cannot establish that a model is scientifically correct or authorized for production. Technical review and application-level validation remain essential.
VMblog: Quantum-HPC integration is moving from roadmap slides to reference deployments, with low-latency links between QPUs and GPU nodes. What will you show at SC26 that is real today, and what’s your honest timeline for hybrid workflows delivering results classical systems can’t?
Fried: Today, teams can develop quantum applications in Classiq, synthesize circuits for a target backend and connect them to GPU simulation or available quantum hardware. The NVIDIA, Oracle and AMD examples demonstrate working development and simulation workflows. They help teams investigate algorithms, integration overhead and scaling before committing to larger quantum experiments.
The timing of useful quantum advantage will depend on the application, hardware quality and the efficiency of the full implementation. A convincing result must account for data preparation, classical processing, communication and error correction, and compare against a strong classical method at equivalent solution quality. We do not put a single date on that across industries.
Our Fault Tolerance Engine helps make the discussion more concrete. It translates logical programs into fault-tolerant execution plans and estimates their resource and reliability requirements. Its current surface-code implementation allows teams to explore how layout, routing and hardware noise affect whether an application could run successfully on a future machine.
VMblog: How do sovereign AI, data residency, and geopolitical risk influence your product design and go-to-market, and how do you support on-prem, cloud, neocloud, and hybrid deployments?
Fried: Where a quantum workflow runs is becoming an important procurement decision. Our multi-region cloud approach is relevant to organizations seeking regional deployment options, while our connections to different compute providers support flexibility in where applications are tested and executed.
For an HPC operator, the practical requirement is to understand the location of each part of the workflow, including development services, stored code, simulation and QPU execution. Selecting a regional software service should be accompanied by a review of the downstream execution providers and their data handling.
We approach deployment discussions around the customer’s requirements for data location, access and integration with existing infrastructure. Cloud, dedicated and on-premises requirements need to be scoped to the specific deployment. Keeping application models portable helps organizations adapt as their infrastructure choices and regional requirements evolve.
VMblog: Talent shortages and lock-in fears are real. How are you making quantum infrastructure easier to operate and more portable through open software stacks, containers, Kubernetes integration, or cross-vendor programming models, so lean teams can do more with less?
Fried: We make quantum development more accessible by allowing teams to work at the algorithm level. Qmod and the Python SDK express the intended computation, while synthesis handles much of the detailed circuit construction and optimization. AI assistance and reusable application examples help domain specialists reach a working implementation sooner.
Portability comes from retaining that high-level model and generating implementations for supported targets. Teams can compare hardware choices and optimize for different constraints without rebuilding the application from its individual gates. Interoperability with CUDA-Q and circuit formats such as OpenQASM helps connect that work to existing tools.
The OCI demonstration also used the NVIDIA cuQuantum Appliance in Docker with MPI across eight GPUs. That is a concrete example of quantum development using infrastructure familiar to HPC practitioners. The goal is to let a lean team accumulate reusable application knowledge as the available hardware improves.
VMblog: Looking toward SC27 and beyond, which bets are you making? What will attendees be talking about a year from now that they aren’t talking about today?
Fried: We are investing in AI-assisted quantum engineering, closer integration with accelerated computing and software that prepares applications for fault-tolerant hardware. The common goal is to make it easier to turn a promising algorithm into an implementation whose performance and resource requirements can be evaluated.
By SC27, we expect more conversations to start with a specific application and ask what combination of logical qubits, physical resources, runtime and reliability it needs. Our Fault Tolerance Engine supports that kind of assessment today, and improvements in algorithms and compilation can change those requirements substantially.
We also expect the quality of the application pipeline to receive more attention. Access to processors becomes more valuable when teams have well-developed algorithms, credible benchmarks and people who understand the underlying problem. Building that capability now gives organizations a stronger basis for using quantum hardware as it matures.
Source links
Intel acquisition announcement
Classiq and NVIDIA CUDA-Q performance result
Oracle and Classiq multi-GPU demonstration
