Product
The Anyon Quantum Supercomputer
A hybrid quantum-classical supercomputer that runs in your data center — superconducting QPUs and NVIDIA GPUs coupled into one low-latency machine, engineered to scale from a single module to a networked fabric.
≤ 20 mK
Operating temperature
NVQLink
Tight GPU coupling
CUDA-Q
Hybrid programming
20 → 1M+
Qubit scaling path
Inside the system
One integrated machine
Anyon engineers the full stack — from the superconducting processor to the GPU it is coupled to — so every layer works as one tightly coupled supercomputer.
Superconducting QPU
The QPU begins as a single chip of parametrically coupled superconducting qubits, then grows into a multi-chiplet processor through advanced packaging. The same architecture is built to scale out across a local entanglement network toward a quantum supercomputer.
Cryogenic platform
A dilution refrigerator holds the QPU below 20 millikelvin — the cold, low-noise environment superconducting qubits require — with modular wiring engineered for cooldown after cooldown.
Quantum control system
Real-time microwave pulse generation, readout, and self-calibration keep every qubit in tune — the classical control layer that drives the quantum hardware at the millikelvin stage.
NVIDIA GPU integration
NVQLink tightly couples the QPU to NVIDIA GPUs, and CUDA-Q programs both as one machine — built for low-latency hybrid quantum-classical workloads and real-time quantum error correction.
How it scales
Scale up, scale out, integrate
The same architecture grows from a single chiplet to a data-center-spanning fabric. Explore the MQN architecture →
Quantum chiplets
More qubits per module by bonding more fab-tolerant chiplets onto a shared routing wafer.
Quantum networking
Microwave quantum links join modules; microwave-to-optical transduction is designed to extend entanglement across data centers.
Quantum-classical coupling
Tight NVQLink + CUDA-Q coupling fuses quantum and classical compute into one hybrid supercomputer.
Specifications
At a glance
- Qubit modality
- Superconducting qubits
- Entangling gates
- Parametric coupling — fabrication-tolerant
- Operating temperature
- ≤ 20 millikelvin (dilution refrigeration)
- Classical integration
- NVIDIA GPU via NVQLink · CUDA-Q hybrid runtime
- Quantum networking
- Microwave quantum links · microwave-to-optical transduction (cross-site)
- Architecture
- Modular Quantum Network (MQN): chiplet → module → networked data center
- Control
- Self-calibrating Quantum Control System (real-time pulse + readout)
- Deployment
- Commercial data center / on-premises
- Scale
- From 20 qubits toward 1M+ across the fabric — 49 per chiplet, 1,225+ per module by design (capability)
Deployment
Deployed in your data center
The Anyon supercomputer is delivered on-premises — installed in your own data center and integrated with your existing HPC and GPU infrastructure.
On-premises
A tightly coupled quantum-classical system installed in your data center — the complete Anyon system, integrated with your HPC and GPU infrastructure.
Under your control
The system runs entirely on your own infrastructure, behind your firewall — your data and workloads stay in your data center, with no third-party cloud in the loop.
No usage fees
The machine is yours — no metered quantum-cloud API charges and no availability limits. Run experiments without restriction and push application research as far as you need.
Bring quantum into your computing roadmap
Talk to our team about deploying a hybrid quantum-classical supercomputer on-premises, in your own data center.