The k&z Quantum Supercomputer platform represents the most advanced commercially available quantum computing infrastructure on the market today. Built around our proprietary QPU Block architecture, each system delivers thousands of superconducting transmon qubits organized into modular, fault-tolerant blocks that can be reserved independently or combined into larger logical clusters. Every QPU Block operates inside a custom-engineered dilution refrigerator cooled to approximately 15 millikelvin — a temperature colder than interstellar space — where quantum coherence times exceed 300 microseconds on average, enabling deep circuit execution with meaningful algorithmic advantage.
Unlike commodity cloud quantum offerings that multiplex workloads across shared QPU resources, k&z Quantum Supercomputers give research teams dedicated access to physical qubit planes during their reservation window. This eliminates crosstalk from neighboring tenants, ensures reproducible calibration baselines, and provides the deterministic scheduling guarantees that serious computational research demands. Whether you are training variational quantum eigensolvers for molecular simulation, executing Grover-accelerated search across combinatorial landscapes, or running quantum approximate optimization algorithms on logistics networks with millions of constraints, QPU Blocks deliver the raw coherent compute fabric you need.
Each QPU Block ships with k&z's proprietary Meridian error-mitigation stack, which applies zero-noise extrapolation, probabilistic error cancellation, and dynamical decoupling sequences in real time during circuit execution. Meridian reduces effective error rates by up to two orders of magnitude compared to raw hardware performance, enabling circuits of 2,000+ two-qubit gate depth to return statistically meaningful results. The Meridian stack is transparent to the user — it runs automatically at the firmware level and requires no changes to your quantum programs. For teams that wish to implement custom mitigation strategies, Meridian can be partially or fully disabled via the k&z Control API.
The block reservation model is designed for flexibility. Teams can reserve as few as one QPU Block (256 physical qubits) for exploratory work or scale to 16 contiguous blocks (4,096 physical qubits) for large-scale quantum simulations. Reservations are available in one-hour, eight-hour, daily, weekly, and monthly increments, with volume discounts applied automatically at the weekly tier and above. All reservations include unlimited shot execution — there are no per-shot fees, no throttling, and no queue contention. When your block is reserved, it is yours exclusively.
Key Capabilities
Modular QPU Block Architecture
Each QPU Block contains 256 physical superconducting transmon qubits arranged in a heavy-hex lattice topology. Blocks interconnect via cryogenic microwave links operating at 6-8 GHz, enabling multi-block entanglement with cross-block two-qubit gate fidelities exceeding 99.2%. Reserve one block for focused experiments or combine up to 16 blocks for computations requiring thousands of qubits.
Meridian Error Mitigation
Our firmware-level Meridian stack combines zero-noise extrapolation (ZNE), probabilistic error cancellation (PEC), and tailored dynamical decoupling to suppress decoherence and gate errors in real time. Meridian extends effective circuit depth by 10-50x compared to unmitigated execution, making near-term quantum advantage achievable on today's hardware without waiting for full fault tolerance.
Deterministic Scheduling
Unlike queue-based cloud platforms where job submission can mean hours of unpredictable wait time, k&z QPU Block reservations guarantee immediate, exclusive hardware access. Submit circuits and receive results in real time with sub-second job-dispatch latency. This deterministic model is critical for iterative algorithms like VQE, QAOA, and quantum machine learning training loops.
Integrated Classical Control Plane
Every QPU Block is paired with a dedicated classical control rack housing custom FPGA-based pulse sequencers, arbitrary waveform generators, and real-time readout discriminators. The control plane executes pulse-level programs with nanosecond timing precision and supports mid-circuit measurement, conditional logic, and dynamic circuit construction without round-trip latency to external servers.
Cryo-Optical Interconnect Fabric
Multi-block configurations leverage k&z's proprietary cryo-optical interconnect, which converts microwave qubit signals to optical photons for low-loss transmission between dilution refrigerators. This fabric enables distributed quantum computing across physically separate cryostats while maintaining entanglement fidelities above 98.5%, a capability no other commercial provider offers at scale.
Full-Stack API Access
Program at the level that suits your team — from high-level Qiskit and Cirq circuit descriptions, through our intermediate QASM-Z representation, down to raw pulse-level control via the k&z Pulse SDK. The k&z Control API exposes every tunable parameter of the hardware, including qubit frequencies, coupling strengths, readout integration windows, and dynamical decoupling sequences.
Technical Specifications
| Parameter | Specification |
|---|---|
| Qubits per QPU Block | 256 superconducting transmon qubits (heavy-hex lattice) |
| Maximum Cluster Size | 16 contiguous blocks (4,096 physical qubits) |
| Single-Qubit Gate Fidelity | ≥ 99.95% (randomized benchmarking) |
| Two-Qubit Gate Fidelity (Intra-Block) | ≥ 99.7% (cross-entropy benchmarking) |
| Two-Qubit Gate Fidelity (Cross-Block) | ≥ 99.2% (cryo-optical interconnect) |
| T1 Coherence Time | 300–500 μs (median across block) |
| T2 Coherence Time | 200–400 μs (echo-based measurement) |
| Readout Fidelity | ≥ 99.5% (single-shot dispersive readout) |
| Operating Temperature | ~15 mK (dilution refrigerator base stage) |
| Mid-Circuit Measurement | Supported (feed-forward latency < 500 ns) |
| Gate Clock Speed | Single-qubit: 25 ns | Two-qubit (CZ): 60 ns |
| Reservation Granularity | 1 hour, 8 hours, daily, weekly, monthly |
| Shot Execution | Unlimited during reservation (no per-shot fees) |
| SDK Compatibility | Qiskit, Cirq, Pennylane, QASM 3.0, k&z Pulse SDK |
| Calibration Cadence | Automated every 4 hours; on-demand recalibration available |
Ideal For
- Quantum chemistry and materials science research — Run variational quantum eigensolver (VQE) and quantum phase estimation (QPE) algorithms to compute molecular ground-state energies, reaction pathways, and electronic structure properties for drug discovery and advanced materials design.
- Combinatorial optimization at scale — Deploy quantum approximate optimization algorithms (QAOA) and quantum annealing-inspired routines across logistics, supply chain, portfolio optimization, and scheduling problems with millions of variables and constraints.
- Quantum machine learning research — Train parameterized quantum circuits, quantum kernel methods, and hybrid quantum-classical neural networks on real hardware with the fast iteration cycles that reproducible, dedicated access enables.
- Quantum error correction development — Implement and benchmark surface codes, color codes, and novel QEC schemes on large qubit arrays with the mid-circuit measurement and feed-forward capabilities required for real-time syndrome extraction and decoding.
- National laboratory and academic research programs — Support multi-year quantum computing research programs with predictable capacity planning, institutional pricing, and integration with existing HPC workflows via k&z's hybrid orchestration layer.
- Benchmarking and algorithmic development — Characterize hardware performance, develop new quantum algorithms, and publish reproducible results using hardware that provides consistent, well-calibrated baselines across extended experimental campaigns.
The k&z Quantum Supercomputer platform is engineered for researchers and engineers who refuse to compromise on hardware quality, access predictability, or scientific reproducibility. Every QPU Block undergoes rigorous acceptance testing — including full tomographic characterization of all single-qubit and two-qubit gates, T1/T2 coherence mapping across the entire qubit plane, and crosstalk matrix measurement — before it is made available for reservation. Calibration data, error maps, and hardware health metrics are published continuously to your dashboard, giving you complete visibility into the physical state of your reserved quantum resources.
Getting started is straightforward. Contact our quantum solutions team to discuss your computational requirements, preferred block count, and reservation schedule. We will provision your QPU Blocks, configure your access credentials, and deliver a dedicated onboarding session with our quantum applications engineers. From first contact to first circuit execution, most teams are running production workloads within 48 hours.