Swiss IBM Quantum System Two puts shared research access ahead of qubit counts

IBM and Lockheed Martin said on September 10, 2026, that they are establishing a Swiss quantum innovation hub at ETH Zurich, anchored by Switzerland’s first IBM Quantum System Two at the Swiss National Supercomputing Centre, or CSCS, in Lugano.¹ The system is expected to be operational by the end of 2026. It will be operated by IBM and use IBM’s Nighthawk r2 processor, which IBM describes as a 120-programmable-qubit chip capable of running more than 100,000 circuits per second.²

The practical significance is not simply that Switzerland is getting a national quantum computer. It is that the machine is being deployed as shared research infrastructure. ETH Zurich will coordinate access for Swiss academia, startups and industry, while CSCS provides the technical environment alongside one of Europe’s most visible classical supercomputing systems.³ That makes the project a test of a maturing question in quantum computing: how to design useful workloads for noisy quantum processors before fully fault-tolerant quantum computers exist.

Who gets access, and when

Access will be mediated through ETH Zurich rather than opened as an unrestricted public service. IBM says organizations that join the hub through ETH Zurich will be able to use IBM’s cloud-accessible quantum fleet immediately, then gain access to Switzerland’s dedicated IBM Quantum System Two after deployment at CSCS.¹ ETH Zurich says it will coordinate use for researchers in the ETH Domain, other Swiss universities and selected Swiss companies.³

That access model matters. In practice, “having a quantum computer” does not mean every researcher gets a terminal directly connected to a quantum processing unit. Users will submit workloads through IBM’s quantum software and cloud environment, with access rights, scheduling and resource allocation coordinated by ETH Zurich and CSCS. IBM Switzerland’s local release says IBM will install, maintain and operate the computer, while ETH Zurich and CSCS will provide power, cooling, security and other infrastructure, and manage allocation of available computing resources.⁴

The agreement is initially planned through 2029, giving Swiss institutions a three-year window to develop use cases, train users and benchmark hybrid workflows on dedicated hardware.⁴ During installation and throughout the agreement, ETH Zurich users will also have cloud access to IBM quantum systems and related resources. That should allow teams to begin porting code, testing circuits and building workflow templates before the Lugano system enters service.⁴

Why CSCS is central to the deployment

The location is as important as the hardware. ETH Zurich says the IBM Quantum System Two will be installed in the same facility that houses Alps, CSCS’s AI-enabled supercomputer.³ That colocation points to a quantum-classical computing model in which quantum processors act as accelerators for specific subroutines, not replacements for high-performance computing.

IBM Quantum System Two is built for this kind of integration. ETH Zurich describes it as an architecture combining cryogenic infrastructure, classical servers for hybrid workflows and modular control electronics for quantum processing units.³ IBM similarly frames System Two as part of “quantum-centric supercomputing,” where quantum hardware is surrounded by classical runtime infrastructure that prepares circuits, processes measurement results, applies error mitigation and iterates algorithms.²

For researchers, the workflow will look more like heterogeneous computing than a standalone quantum experiment. A classical HPC system can generate molecular Hamiltonians, optimize parameters, simulate smaller circuit instances, precondition data or analyze large batches of measurement results. The quantum processor runs circuits that are hard or costly to emulate classically, then returns samples or expectation values to the classical workflow. The classical system updates the next circuit batch, evaluates convergence and handles data management.

That loop is especially relevant for current quantum processors, which remain noisy and probabilistic. Throughput — the number of circuits that can be executed and measured per second — can be as important as qubit count because near-term algorithms often require many repeated circuit executions to estimate observables, sweep parameters or compare error-mitigation strategies. IBM says Nighthawk r2 can execute more than 100,000 circuits per second and has demonstrated accurate computations on circuits with 7,500 gates.²

What researchers can realistically do before fault tolerance

The system should not be read as a general-purpose, fault-tolerant quantum computer. Fault tolerance requires error-corrected logical qubits that can run long computations reliably despite physical qubit errors. The announced Swiss system is instead a high-end, near-term research platform for designing, testing and benchmarking workloads on noisy quantum hardware.

The feasible experiments fall into several categories.

First, chemistry and materials science teams can explore quantum simulations of molecules, correlated materials and reaction processes that are difficult to model exactly with classical methods. ETH Zurich highlights materials, molecules and chemical processes as target areas where the combination of quantum and classical systems could open new research paths.³ Near-term work is likely to focus on reduced models, active-space chemistry, variational eigensolvers, time-evolution experiments and comparisons against classical approximations, rather than full-scale industrial molecule discovery.

Second, algorithm researchers can test workload design itself: circuit depth, layout choices, transpilation strategies, error mitigation, qubit reset, measurement grouping and batching. This is where Nighthawk r2’s circuit throughput could matter. More circuit executions allow users to test more parameter settings, gather more samples and evaluate whether an algorithm is robust enough to survive device noise.²

Third, optimization and financial-services researchers can benchmark quantum and hybrid heuristics on structured problems. IBM lists optimization and financial services among the areas the hub is intended to support.¹ In the near term, these experiments are best understood as algorithmic research and comparative benchmarking, not guaranteed production advantage. Useful outputs may include identifying problem structures that map well to quantum circuits, developing hybrid solvers and building evidence about when quantum resources improve a broader workflow.

Fourth, the hub will support workforce development and application engineering. IBM says the initiative includes access to the IBM Quantum Network and IBM Quantum Platform learning offerings, including coursework, certifications, workshops and events.¹ For a national research ecosystem, that training layer is not secondary. Quantum computing still requires specialists who understand both the physics of hardware constraints and the software engineering of hybrid workflows.

Lockheed Martin’s role and the defense-backed context

The hub is being created by IBM and Lockheed Martin through an offset agreement with armasuisse, Switzerland’s Federal Office for Defence Procurement.¹ PR Newswire distributed the announcement at 03:00 ET on September 10, 2026, confirming the formal timing and the companies’ framing of the Swiss hub as an extension of their ongoing collaboration.⁵

Lockheed Martin and IBM plan two initial joint development projects: one on quantum sensing applications for navigation and another on improving additive manufacturing of metallic alloys.¹ Independent defense-industry coverage emphasized the same context, describing the project as a defense-backed push involving navigation, AI and advanced manufacturing, with ETH Zurich managing access for Swiss companies, startups and academic institutions.⁶

Those projects show how the hub may connect quantum computing to adjacent advanced-computing domains. Quantum sensing is not the same as gate-based quantum computing, but navigation applications can involve quantum physics, signal processing, optimization and simulation. Additive manufacturing of metallic alloys may require classical simulation, materials modeling, AI and potentially quantum chemistry subroutines. The common thread is not a single quantum application, but a shared infrastructure stack for testing whether quantum processors can improve parts of larger R&D workflows.

The real benchmark

For advanced computing readers, the Swiss deployment should be evaluated less as a qubit-count milestone than as an infrastructure experiment. The important questions are operational: who receives allocations, how easily quantum jobs can be embedded into CSCS-style workflows, what software abstractions emerge, and whether researchers can define workloads that extract measurable value from noisy hardware.

If the hub succeeds, its output may not be one dramatic “killer app.” It may be a portfolio of better circuit benchmarks, hybrid algorithms, trained users, reproducible workflow patterns and domain-specific evidence about where quantum computing helps. That would still be meaningful. Before fault tolerance arrives, the frontier is not only building larger quantum processors; it is learning how to use them as part of a serious computing environment.

Swiss IBM Quantum System Two puts shared research access ahead of qubit counts

More from Siften Technology

Apple’s iPhone Duo tests whether iOS is ready for foldables

Things to know

Comments

Source Material

Who gets access, and when

Why CSCS is central to the deployment

What researchers can realistically do before fault tolerance

Lockheed Martin’s role and the defense-backed context

The real benchmark

Key topics covered

CISA Confirms Ransomware Exploitation of WatchGuard Firebox RCE Flaw

CoreWeave’s Physical AI Push Shows GPUs Are No Longer Enough

NASA and IBM release open lunar AI model for reproducible Moon science