Alice & Bob, Hyperion say quantum’s value in AI will hinge on the full stack

Alice & Bob and Hyperion Research released a report Tuesday arguing that HPC centers will only get value from fault-tolerant quantum computing if software, workflows and operations mature alongside hardware. The findings, based on 15 interviews with supercomputing practitioners across Europe, APAC and the U.S., warn that deploying immature systems too early could damage trust in quantum computing.

Why it matters: - HPC centers are starting to plan for fault-tolerant quantum computing, but the report says hardware alone will not deliver usable AI and HPC gains. - The report argues that software readiness, workflow integration and operational expertise will determine whether quantum systems add value in real workloads. - An immature stack deployed too early could undermine trust in quantum computing across the HPC community for years.

What happened: - Alice & Bob and Hyperion Research released "HPC-Quantum-AI: Shaping the Next Compute Era" on Tuesday. - The report focuses on what HPC-Quantum-AI systems will need as quantum hardware moves toward fault tolerance. - The findings are based on 15 interviews with directors, quantum program leads and engineers at leading supercomputing institutions in Europe, APAC and the U.S. - Interviewees came from Argonne National Laboratory, Lawrence Berkeley National Laboratory/NERSC, Oak Ridge National Laboratory, RIKEN, Munich Quantum Valley and Pawsey Supercomputing Centre.

The details: - The report says the key question for HPC centers is not whether qubits will arrive, but how to use them effectively. - Juliette Peyronnet, GM US at Alice & Bob, said quantum processors will matter in HPC only if they can execute applications. - The report identifies three near-term actions for the quantum-HPC community. - One action is to map the quantum-HPC-AI design space by measuring how much latency a workload can tolerate. - The report says that answer cannot be bought and must be measured through testbeds. - Another action is to unify the software stack around open standards for layers that do not depend on qubit modality, including device management and job submission. - The report says compilers and programming models should remain open to competition. - A third action is to co-design workflows so HPC, AI and quantum teams work on the same problem. - The report cites a quantum chemistry workflow targeting a 10,000-fold speed-up. - The report closes by making the case for on-premise deployment, saying integration expertise builds inside centers that operate a quantum machine with full access. - The report argues the first fault-tolerant machines will go to centers already building that readiness. - This is the second report in Alice & Bob's "Seizing Quantum's Edge" series on HPC-Quantum-AI integration. - The 2025 edition said HPC centers should prepare for early fault-tolerant quantum computing.

Between the lines: - The report is shifting the conversation from quantum hardware capability to system integration and deployment readiness. - That framing favors institutions that can run testbeds, build operational know-how and shape standards before large-scale commercial rollout. - Hyperion Research's participation signals that the quantum opportunity is being treated as an HPC market-planning issue, not just a research milestone. - Bob Sorensen, SVP of Research and Chief Analyst for Quantum Computing at Hyperion Research, said the report is timely because end users need guidance on integrating quantum into current and future advanced computing ecosystems.

What's next: - HPC centers are expected to keep building testbeds and evaluating latency-sensitive workflows. - Quantum vendors and centers will likely face pressure to align on open standards for the non-qubit parts of the stack. - The report suggests early fault-tolerant systems will land first at centers already prepared to operate them on-site. - The full report is available here.